diff --git a/sql/upgrades/131.sql b/sql/upgrades/131.sql new file mode 100644 index 0000000..b57d027 --- /dev/null +++ b/sql/upgrades/131.sql @@ -0,0 +1,223 @@ +-- SWH Indexer DB schema upgrade +-- from_version: 130 +-- to_version: 131 +-- description: _add function returns the inserted rows + +insert into dbversion(version, release, description) + values(131, now(), 'Work In Progress'); + +drop function swh_content_mimetype_add(boolean); +drop function swh_content_language_add(boolean); +drop function swh_content_ctags_add(boolean); +drop function swh_content_fossology_license_add(boolean); +drop function swh_content_metadata_add(boolean); +drop function swh_revision_intrinsic_metadata_add(boolean); +drop function swh_origin_intrinsic_metadata_add(boolean); + +create or replace function swh_content_mimetype_add(conflict_update boolean) + returns bigint + language plpgsql +as $$ +declare + res bigint; +begin + if conflict_update then + insert into content_mimetype (id, mimetype, encoding, indexer_configuration_id) + select id, mimetype, encoding, indexer_configuration_id + from tmp_content_mimetype tcm + on conflict(id, indexer_configuration_id) + do update set mimetype = excluded.mimetype, + encoding = excluded.encoding; + else + insert into content_mimetype (id, mimetype, encoding, indexer_configuration_id) + select id, mimetype, encoding, indexer_configuration_id + from tmp_content_mimetype tcm + on conflict(id, indexer_configuration_id) + do nothing; + end if; + get diagnostics res = ROW_COUNT; + return res; +end +$$; + +comment on function swh_content_mimetype_add(boolean) IS 'Add new content mimetypes'; + +create or replace function swh_content_language_add(conflict_update boolean) + returns bigint + language plpgsql +as $$ +declare + res bigint; +begin + if conflict_update then + insert into content_language (id, lang, indexer_configuration_id) + select id, lang, indexer_configuration_id + from tmp_content_language tcl + on conflict(id, indexer_configuration_id) + do update set lang = excluded.lang; + else + insert into content_language (id, lang, indexer_configuration_id) + select id, lang, indexer_configuration_id + from tmp_content_language tcl + on conflict(id, indexer_configuration_id) + do nothing; + end if; + get diagnostics res = ROW_COUNT; + return res; +end +$$; + +comment on function swh_content_language_add(boolean) IS 'Add new content languages'; + + +create or replace function swh_content_ctags_add(conflict_update boolean) + returns bigint + language plpgsql +as $$ +declare + res bigint; +begin + if conflict_update then + delete from content_ctags + where id in (select tmp.id + from tmp_content_ctags tmp + inner join indexer_configuration i on i.id=tmp.indexer_configuration_id); + end if; + + insert into content_ctags (id, name, kind, line, lang, indexer_configuration_id) + select id, name, kind, line, lang, indexer_configuration_id + from tmp_content_ctags tct + on conflict(id, hash_sha1(name), kind, line, lang, indexer_configuration_id) + do nothing; + get diagnostics res = ROW_COUNT; + return res; +end +$$; + +comment on function swh_content_ctags_add(boolean) IS 'Add new ctags symbols per content'; + +create or replace function swh_content_fossology_license_add(conflict_update boolean) + returns bigint + language plpgsql +as $$ +declare + res bigint; +begin + -- insert unknown licenses first + insert into fossology_license (name) + select distinct license from tmp_content_fossology_license tmp + where not exists (select 1 from fossology_license where name=tmp.license) + on conflict(name) do nothing; + + if conflict_update then + insert into content_fossology_license (id, license_id, indexer_configuration_id) + select tcl.id, + (select id from fossology_license where name = tcl.license) as license, + indexer_configuration_id + from tmp_content_fossology_license tcl + on conflict(id, license_id, indexer_configuration_id) + do update set license_id = excluded.license_id; + end if; + + insert into content_fossology_license (id, license_id, indexer_configuration_id) + select tcl.id, + (select id from fossology_license where name = tcl.license) as license, + indexer_configuration_id + from tmp_content_fossology_license tcl + on conflict(id, license_id, indexer_configuration_id) + do nothing; + + get diagnostics res = ROW_COUNT; + return res; +end +$$; + +create or replace function swh_content_metadata_add(conflict_update boolean) + returns bigint + language plpgsql +as $$ +declare + res bigint; +begin + if conflict_update then + insert into content_metadata (id, metadata, indexer_configuration_id) + select id, metadata, indexer_configuration_id + from tmp_content_metadata tcm + on conflict(id, indexer_configuration_id) + do update set metadata = excluded.metadata; + else + insert into content_metadata (id, metadata, indexer_configuration_id) + select id, metadata, indexer_configuration_id + from tmp_content_metadata tcm + on conflict(id, indexer_configuration_id) + do nothing; + end if; + get diagnostics res = ROW_COUNT; + return res; +end +$$; + +comment on function swh_content_metadata_add(boolean) IS 'Add new content metadata'; + + +create or replace function swh_revision_intrinsic_metadata_add(conflict_update boolean) + returns bigint + language plpgsql +as $$ +declare + res bigint; +begin + if conflict_update then + insert into revision_intrinsic_metadata (id, metadata, mappings, indexer_configuration_id) + select id, metadata, mappings, indexer_configuration_id + from tmp_revision_intrinsic_metadata tcm + on conflict(id, indexer_configuration_id) + do update set + metadata = excluded.metadata, + mappings = excluded.mappings; + else + insert into revision_intrinsic_metadata (id, metadata, mappings, indexer_configuration_id) + select id, metadata, mappings, indexer_configuration_id + from tmp_revision_intrinsic_metadata tcm + on conflict(id, indexer_configuration_id) + do nothing; + end if; + get diagnostics res = ROW_COUNT; + return res; +end +$$; + +comment on function swh_revision_intrinsic_metadata_add(boolean) IS 'Add new revision intrinsic metadata'; + +create or replace function swh_origin_intrinsic_metadata_add( + conflict_update boolean) + returns bigint + language plpgsql +as $$ +declare + res bigint; +begin + perform swh_origin_intrinsic_metadata_compute_tsvector(); + if conflict_update then + insert into origin_intrinsic_metadata (id, metadata, indexer_configuration_id, from_revision, metadata_tsvector, mappings) + select id, metadata, indexer_configuration_id, from_revision, + metadata_tsvector, mappings + from tmp_origin_intrinsic_metadata + on conflict(id, indexer_configuration_id) + do update set + metadata = excluded.metadata, + metadata_tsvector = excluded.metadata_tsvector, + mappings = excluded.mappings, + from_revision = excluded.from_revision; + else + insert into origin_intrinsic_metadata (id, metadata, indexer_configuration_id, from_revision, metadata_tsvector, mappings) + select id, metadata, indexer_configuration_id, from_revision, + metadata_tsvector, mappings + from tmp_origin_intrinsic_metadata + on conflict(id, indexer_configuration_id) + do nothing; + end if; + get diagnostics res = ROW_COUNT; + return res; +end +$$; diff --git a/swh/indexer/sql/30-swh-schema.sql b/swh/indexer/sql/30-swh-schema.sql index eec0f3b..44532be 100644 --- a/swh/indexer/sql/30-swh-schema.sql +++ b/swh/indexer/sql/30-swh-schema.sql @@ -1,145 +1,145 @@ --- --- Software Heritage Indexers Data Model --- -- drop schema if exists swh cascade; -- create schema swh; -- set search_path to swh; create table dbversion ( version int primary key, release timestamptz, description text ); insert into dbversion(version, release, description) - values(130, now(), 'Work In Progress'); + values(131, now(), 'Work In Progress'); -- Computing metadata on sha1's contents -- a SHA1 checksum (not necessarily originating from Git) create domain sha1 as bytea check (length(value) = 20); -- a Git object ID, i.e., a SHA1 checksum create domain sha1_git as bytea check (length(value) = 20); create table indexer_configuration ( id serial not null, tool_name text not null, tool_version text not null, tool_configuration jsonb ); comment on table indexer_configuration is 'Indexer''s configuration version'; comment on column indexer_configuration.id is 'Tool identifier'; comment on column indexer_configuration.tool_version is 'Tool name'; comment on column indexer_configuration.tool_version is 'Tool version'; comment on column indexer_configuration.tool_configuration is 'Tool configuration: command line, flags, etc...'; -- Properties (mimetype, encoding, etc...) create table content_mimetype ( id sha1 not null, mimetype text not null, encoding text not null, indexer_configuration_id bigint not null ); comment on table content_mimetype is 'Metadata associated to a raw content'; comment on column content_mimetype.mimetype is 'Raw content Mimetype'; comment on column content_mimetype.encoding is 'Raw content encoding'; comment on column content_mimetype.indexer_configuration_id is 'Tool used to compute the information'; -- Language metadata create table content_language ( id sha1 not null, lang languages not null, indexer_configuration_id bigint not null ); comment on table content_language is 'Language information on a raw content'; comment on column content_language.lang is 'Language information'; comment on column content_language.indexer_configuration_id is 'Tool used to compute the information'; -- ctags information per content create table content_ctags ( id sha1 not null, name text not null, kind text not null, line bigint not null, lang ctags_languages not null, indexer_configuration_id bigint not null ); comment on table content_ctags is 'Ctags information on a raw content'; comment on column content_ctags.id is 'Content identifier'; comment on column content_ctags.name is 'Symbol name'; comment on column content_ctags.kind is 'Symbol kind (function, class, variable, const...)'; comment on column content_ctags.line is 'Symbol line'; comment on column content_ctags.lang is 'Language information for that content'; comment on column content_ctags.indexer_configuration_id is 'Tool used to compute the information'; create table fossology_license( id smallserial, name text not null ); comment on table fossology_license is 'Possible license recognized by license indexer'; comment on column fossology_license.id is 'License identifier'; comment on column fossology_license.name is 'License name'; create table content_fossology_license ( id sha1 not null, license_id smallserial not null, indexer_configuration_id bigint not null ); comment on table content_fossology_license is 'license associated to a raw content'; comment on column content_fossology_license.id is 'Raw content identifier'; comment on column content_fossology_license.license_id is 'One of the content''s license identifier'; comment on column content_fossology_license.indexer_configuration_id is 'Tool used to compute the information'; -- The table content_metadata provides a translation to files -- identified as potentially containning metadata with a translation tool (indexer_configuration_id) create table content_metadata( id sha1 not null, metadata jsonb not null, indexer_configuration_id bigint not null ); comment on table content_metadata is 'metadata semantically translated from a content file'; comment on column content_metadata.id is 'sha1 of content file'; comment on column content_metadata.metadata is 'result of translation with defined format'; comment on column content_metadata.indexer_configuration_id is 'tool used for translation'; -- The table revision_intrinsic_metadata provides a minimal set of intrinsic -- metadata detected with the detection tool (indexer_configuration_id) and -- aggregated from the content_metadata translation. create table revision_intrinsic_metadata( id sha1_git not null, metadata jsonb not null, indexer_configuration_id bigint not null, mappings text array not null ); comment on table revision_intrinsic_metadata is 'metadata semantically detected and translated in a revision'; comment on column revision_intrinsic_metadata.id is 'sha1_git of revision'; comment on column revision_intrinsic_metadata.metadata is 'result of detection and translation with defined format'; comment on column revision_intrinsic_metadata.indexer_configuration_id is 'tool used for detection'; comment on column revision_intrinsic_metadata.mappings is 'type of metadata files used to obtain this metadata (eg. pkg-info, npm)'; create table origin_intrinsic_metadata( id text not null, -- origin url metadata jsonb, indexer_configuration_id bigint not null, from_revision sha1_git not null, metadata_tsvector tsvector, mappings text array not null ); comment on table origin_intrinsic_metadata is 'keeps intrinsic metadata for an origin'; comment on column origin_intrinsic_metadata.id is 'url of the origin'; comment on column origin_intrinsic_metadata.metadata is 'metadata extracted from a revision'; comment on column origin_intrinsic_metadata.indexer_configuration_id is 'tool used to generate this metadata'; comment on column origin_intrinsic_metadata.from_revision is 'sha1 of the revision this metadata was copied from.'; comment on column origin_intrinsic_metadata.mappings is 'type of metadata files used to obtain this metadata (eg. pkg-info, npm)'; diff --git a/swh/indexer/sql/40-swh-func.sql b/swh/indexer/sql/40-swh-func.sql index 1b325ec..81fca5b 100644 --- a/swh/indexer/sql/40-swh-func.sql +++ b/swh/indexer/sql/40-swh-func.sql @@ -1,455 +1,472 @@ -- Postgresql index helper function create or replace function hash_sha1(text) returns text language sql strict immutable as $$ select encode(public.digest($1, 'sha1'), 'hex') $$; comment on function hash_sha1(text) is 'Compute sha1 hash as text'; -- create a temporary table called tmp_TBLNAME, mimicking existing table -- TBLNAME -- -- Args: -- tblname: name of the table to mimic create or replace function swh_mktemp(tblname regclass) returns void language plpgsql as $$ begin execute format(' create temporary table if not exists tmp_%1$I (like %1$I including defaults) on commit delete rows; alter table tmp_%1$I drop column if exists object_id; ', tblname); return; end $$; -- create a temporary table for content_mimetype tmp_content_mimetype, create or replace function swh_mktemp_content_mimetype() returns void language sql as $$ create temporary table if not exists tmp_content_mimetype ( like content_mimetype including defaults ) on commit delete rows; $$; comment on function swh_mktemp_content_mimetype() IS 'Helper table to add mimetype information'; -- add tmp_content_mimetype entries to content_mimetype, overwriting -- duplicates if conflict_update is true, skipping duplicates otherwise. -- -- If filtering duplicates is in order, the call to -- swh_content_mimetype_missing must take place before calling this -- function. -- --- -- operates in bulk: 0. swh_mktemp(content_mimetype), 1. COPY to tmp_content_mimetype, -- 2. call this function create or replace function swh_content_mimetype_add(conflict_update boolean) - returns void + returns bigint language plpgsql as $$ +declare + res bigint; begin if conflict_update then insert into content_mimetype (id, mimetype, encoding, indexer_configuration_id) select id, mimetype, encoding, indexer_configuration_id from tmp_content_mimetype tcm - on conflict(id, indexer_configuration_id) - do update set mimetype = excluded.mimetype, - encoding = excluded.encoding; - + on conflict(id, indexer_configuration_id) + do update set mimetype = excluded.mimetype, + encoding = excluded.encoding; else insert into content_mimetype (id, mimetype, encoding, indexer_configuration_id) select id, mimetype, encoding, indexer_configuration_id from tmp_content_mimetype tcm - on conflict(id, indexer_configuration_id) do nothing; + on conflict(id, indexer_configuration_id) + do nothing; end if; - return; + get diagnostics res = ROW_COUNT; + return res; end $$; comment on function swh_content_mimetype_add(boolean) IS 'Add new content mimetypes'; -- add tmp_content_language entries to content_language, overwriting -- duplicates if conflict_update is true, skipping duplicates otherwise. -- -- If filtering duplicates is in order, the call to -- swh_content_language_missing must take place before calling this -- function. -- -- operates in bulk: 0. swh_mktemp(content_language), 1. COPY to -- tmp_content_language, 2. call this function create or replace function swh_content_language_add(conflict_update boolean) - returns void + returns bigint language plpgsql as $$ +declare + res bigint; begin if conflict_update then insert into content_language (id, lang, indexer_configuration_id) select id, lang, indexer_configuration_id - from tmp_content_language tcl - on conflict(id, indexer_configuration_id) - do update set lang = excluded.lang; - + from tmp_content_language tcl + on conflict(id, indexer_configuration_id) + do update set lang = excluded.lang; else insert into content_language (id, lang, indexer_configuration_id) select id, lang, indexer_configuration_id - from tmp_content_language tcl - on conflict(id, indexer_configuration_id) - do nothing; + from tmp_content_language tcl + on conflict(id, indexer_configuration_id) + do nothing; end if; - return; + get diagnostics res = ROW_COUNT; + return res; end $$; comment on function swh_content_language_add(boolean) IS 'Add new content languages'; -- create a temporary table for retrieving content_language create or replace function swh_mktemp_content_language() returns void language sql as $$ create temporary table if not exists tmp_content_language ( like content_language including defaults ) on commit delete rows; $$; comment on function swh_mktemp_content_language() is 'Helper table to add content language'; -- create a temporary table for content_ctags tmp_content_ctags, create or replace function swh_mktemp_content_ctags() returns void language sql as $$ create temporary table if not exists tmp_content_ctags ( like content_ctags including defaults ) on commit delete rows; $$; comment on function swh_mktemp_content_ctags() is 'Helper table to add content ctags'; -- add tmp_content_ctags entries to content_ctags, overwriting -- duplicates if conflict_update is true, skipping duplicates otherwise. -- -- operates in bulk: 0. swh_mktemp(content_ctags), 1. COPY to tmp_content_ctags, -- 2. call this function create or replace function swh_content_ctags_add(conflict_update boolean) - returns void + returns bigint language plpgsql as $$ +declare + res bigint; begin if conflict_update then delete from content_ctags where id in (select tmp.id from tmp_content_ctags tmp inner join indexer_configuration i on i.id=tmp.indexer_configuration_id); end if; insert into content_ctags (id, name, kind, line, lang, indexer_configuration_id) select id, name, kind, line, lang, indexer_configuration_id from tmp_content_ctags tct - on conflict(id, hash_sha1(name), kind, line, lang, indexer_configuration_id) - do nothing; - return; + on conflict(id, hash_sha1(name), kind, line, lang, indexer_configuration_id) + do nothing; + get diagnostics res = ROW_COUNT; + return res; end $$; comment on function swh_content_ctags_add(boolean) IS 'Add new ctags symbols per content'; create type content_ctags_signature as ( id sha1, name text, kind text, line bigint, lang ctags_languages, tool_id integer, tool_name text, tool_version text, tool_configuration jsonb ); -- Search within ctags content. -- create or replace function swh_content_ctags_search( expression text, l integer default 10, last_sha1 sha1 default '\x0000000000000000000000000000000000000000') returns setof content_ctags_signature language sql as $$ select c.id, name, kind, line, lang, i.id as tool_id, tool_name, tool_version, tool_configuration from content_ctags c inner join indexer_configuration i on i.id = c.indexer_configuration_id where hash_sha1(name) = hash_sha1(expression) and c.id > last_sha1 order by id limit l; $$; comment on function swh_content_ctags_search(text, integer, sha1) IS 'Equality search through ctags'' symbols'; -- create a temporary table for content_fossology_license tmp_content_fossology_license, create or replace function swh_mktemp_content_fossology_license() returns void language sql as $$ create temporary table if not exists tmp_content_fossology_license ( id sha1, license text, indexer_configuration_id integer ) on commit delete rows; $$; comment on function swh_mktemp_content_fossology_license() is 'Helper table to add content license'; -- add tmp_content_fossology_license entries to content_fossology_license, overwriting -- duplicates if conflict_update is true, skipping duplicates otherwise. -- -- operates in bulk: 0. swh_mktemp(content_fossology_license), 1. COPY to -- tmp_content_fossology_license, 2. call this function create or replace function swh_content_fossology_license_add(conflict_update boolean) - returns void - language plpgsql + returns bigint + language plpgsql as $$ +declare + res bigint; begin -- insert unknown licenses first insert into fossology_license (name) select distinct license from tmp_content_fossology_license tmp where not exists (select 1 from fossology_license where name=tmp.license) on conflict(name) do nothing; if conflict_update then insert into content_fossology_license (id, license_id, indexer_configuration_id) select tcl.id, (select id from fossology_license where name = tcl.license) as license, indexer_configuration_id from tmp_content_fossology_license tcl - on conflict(id, license_id, indexer_configuration_id) - do update set license_id = excluded.license_id; - return; + on conflict(id, license_id, indexer_configuration_id) + do update set license_id = excluded.license_id; end if; insert into content_fossology_license (id, license_id, indexer_configuration_id) select tcl.id, (select id from fossology_license where name = tcl.license) as license, indexer_configuration_id from tmp_content_fossology_license tcl - on conflict(id, license_id, indexer_configuration_id) - do nothing; - return; + on conflict(id, license_id, indexer_configuration_id) + do nothing; + + get diagnostics res = ROW_COUNT; + return res; end $$; comment on function swh_content_fossology_license_add(boolean) IS 'Add new content licenses'; + -- content_metadata functions -- add tmp_content_metadata entries to content_metadata, overwriting -- duplicates if conflict_update is true, skipping duplicates otherwise. -- -- If filtering duplicates is in order, the call to -- swh_content_metadata_missing must take place before calling this -- function. -- -- operates in bulk: 0. swh_mktemp(content_language), 1. COPY to -- tmp_content_metadata, 2. call this function create or replace function swh_content_metadata_add(conflict_update boolean) - returns void + returns bigint language plpgsql as $$ +declare + res bigint; begin if conflict_update then insert into content_metadata (id, metadata, indexer_configuration_id) select id, metadata, indexer_configuration_id - from tmp_content_metadata tcm - on conflict(id, indexer_configuration_id) - do update set metadata = excluded.metadata; - + from tmp_content_metadata tcm + on conflict(id, indexer_configuration_id) + do update set metadata = excluded.metadata; else insert into content_metadata (id, metadata, indexer_configuration_id) select id, metadata, indexer_configuration_id from tmp_content_metadata tcm - on conflict(id, indexer_configuration_id) - do nothing; + on conflict(id, indexer_configuration_id) + do nothing; end if; - return; + get diagnostics res = ROW_COUNT; + return res; end $$; comment on function swh_content_metadata_add(boolean) IS 'Add new content metadata'; -- create a temporary table for retrieving content_metadata create or replace function swh_mktemp_content_metadata() returns void language sql as $$ create temporary table if not exists tmp_content_metadata ( like content_metadata including defaults ) on commit delete rows; $$; comment on function swh_mktemp_content_metadata() is 'Helper table to add content metadata'; -- end content_metadata functions -- add tmp_revision_intrinsic_metadata entries to revision_intrinsic_metadata, -- overwriting duplicates if conflict_update is true, skipping duplicates -- otherwise. -- -- If filtering duplicates is in order, the call to -- swh_revision_intrinsic_metadata_missing must take place before calling this -- function. -- -- operates in bulk: 0. swh_mktemp(content_language), 1. COPY to -- tmp_revision_intrinsic_metadata, 2. call this function create or replace function swh_revision_intrinsic_metadata_add(conflict_update boolean) - returns void + returns bigint language plpgsql as $$ +declare + res bigint; begin if conflict_update then insert into revision_intrinsic_metadata (id, metadata, mappings, indexer_configuration_id) select id, metadata, mappings, indexer_configuration_id - from tmp_revision_intrinsic_metadata tcm - on conflict(id, indexer_configuration_id) - do update set - metadata = excluded.metadata, - mappings = excluded.mappings; - + from tmp_revision_intrinsic_metadata tcm + on conflict(id, indexer_configuration_id) + do update set + metadata = excluded.metadata, + mappings = excluded.mappings; else insert into revision_intrinsic_metadata (id, metadata, mappings, indexer_configuration_id) select id, metadata, mappings, indexer_configuration_id from tmp_revision_intrinsic_metadata tcm - on conflict(id, indexer_configuration_id) - do nothing; + on conflict(id, indexer_configuration_id) + do nothing; end if; - return; + get diagnostics res = ROW_COUNT; + return res; end $$; comment on function swh_revision_intrinsic_metadata_add(boolean) IS 'Add new revision intrinsic metadata'; -- create a temporary table for retrieving revision_intrinsic_metadata create or replace function swh_mktemp_revision_intrinsic_metadata() returns void language sql as $$ create temporary table if not exists tmp_revision_intrinsic_metadata ( like revision_intrinsic_metadata including defaults ) on commit delete rows; $$; comment on function swh_mktemp_revision_intrinsic_metadata() is 'Helper table to add revision intrinsic metadata'; -- create a temporary table for retrieving origin_intrinsic_metadata create or replace function swh_mktemp_origin_intrinsic_metadata() returns void language sql as $$ create temporary table if not exists tmp_origin_intrinsic_metadata ( like origin_intrinsic_metadata including defaults ) on commit delete rows; $$; comment on function swh_mktemp_origin_intrinsic_metadata() is 'Helper table to add origin intrinsic metadata'; create or replace function swh_mktemp_indexer_configuration() returns void language sql as $$ create temporary table if not exists tmp_indexer_configuration ( like indexer_configuration including defaults ) on commit delete rows; alter table tmp_indexer_configuration drop column if exists id; $$; -- add tmp_indexer_configuration entries to indexer_configuration, -- skipping duplicates if any. -- -- operates in bulk: 0. create temporary tmp_indexer_configuration, 1. COPY to -- it, 2. call this function to insert and filtering out duplicates create or replace function swh_indexer_configuration_add() returns setof indexer_configuration language plpgsql as $$ begin insert into indexer_configuration(tool_name, tool_version, tool_configuration) select tool_name, tool_version, tool_configuration from tmp_indexer_configuration tmp on conflict(tool_name, tool_version, tool_configuration) do nothing; return query select id, tool_name, tool_version, tool_configuration from tmp_indexer_configuration join indexer_configuration using(tool_name, tool_version, tool_configuration); return; end $$; -- add tmp_origin_intrinsic_metadata entries to origin_intrinsic_metadata, -- overwriting duplicates if conflict_update is true, skipping duplicates -- otherwise. -- -- If filtering duplicates is in order, the call to -- swh_origin_intrinsic_metadata_missing must take place before calling this -- function. -- -- operates in bulk: 0. swh_mktemp(content_language), 1. COPY to -- tmp_origin_intrinsic_metadata, 2. call this function create or replace function swh_origin_intrinsic_metadata_add( conflict_update boolean) - returns void + returns bigint language plpgsql as $$ +declare + res bigint; begin perform swh_origin_intrinsic_metadata_compute_tsvector(); if conflict_update then insert into origin_intrinsic_metadata (id, metadata, indexer_configuration_id, from_revision, metadata_tsvector, mappings) select id, metadata, indexer_configuration_id, from_revision, metadata_tsvector, mappings - from tmp_origin_intrinsic_metadata - on conflict(id, indexer_configuration_id) - do update set - metadata = excluded.metadata, - metadata_tsvector = excluded.metadata_tsvector, - mappings = excluded.mappings, - from_revision = excluded.from_revision; - + from tmp_origin_intrinsic_metadata + on conflict(id, indexer_configuration_id) + do update set + metadata = excluded.metadata, + metadata_tsvector = excluded.metadata_tsvector, + mappings = excluded.mappings, + from_revision = excluded.from_revision; else insert into origin_intrinsic_metadata (id, metadata, indexer_configuration_id, from_revision, metadata_tsvector, mappings) select id, metadata, indexer_configuration_id, from_revision, metadata_tsvector, mappings from tmp_origin_intrinsic_metadata - on conflict(id, indexer_configuration_id) - do nothing; + on conflict(id, indexer_configuration_id) + do nothing; end if; - return; + get diagnostics res = ROW_COUNT; + return res; end $$; comment on function swh_origin_intrinsic_metadata_add(boolean) IS 'Add new origin intrinsic metadata'; -- Compute the metadata_tsvector column in tmp_origin_intrinsic_metadata. -- -- It uses the "pg_catalog.simple" dictionary, as it has no stopword, -- so it should be suitable for proper names and non-English text. create or replace function swh_origin_intrinsic_metadata_compute_tsvector() returns void language plpgsql as $$ begin update tmp_origin_intrinsic_metadata set metadata_tsvector = to_tsvector('pg_catalog.simple', metadata); end $$; diff --git a/swh/indexer/storage/__init__.py b/swh/indexer/storage/__init__.py index d853f17..aa41126 100644 --- a/swh/indexer/storage/__init__.py +++ b/swh/indexer/storage/__init__.py @@ -1,455 +1,500 @@ # Copyright (C) 2015-2020 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import json import psycopg2 import psycopg2.pool from collections import defaultdict, Counter +from typing import Dict, List from swh.storage.common import db_transaction_generator, db_transaction from swh.storage.exc import StorageDBError from . import converters from .db import Db from .exc import IndexerStorageArgumentException, DuplicateId INDEXER_CFG_KEY = 'indexer_storage' MAPPING_NAMES = ['codemeta', 'gemspec', 'maven', 'npm', 'pkg-info'] def get_indexer_storage(cls, args): """Get an indexer storage object of class `storage_class` with arguments `storage_args`. Args: cls (str): storage's class, either 'local' or 'remote' args (dict): dictionary of arguments passed to the storage class constructor Returns: an instance of swh.indexer's storage (either local or remote) Raises: ValueError if passed an unknown storage class. """ if cls == 'remote': from .api.client import RemoteStorage as IndexerStorage elif cls == 'local': from . import IndexerStorage elif cls == 'memory': from .in_memory import IndexerStorage else: raise ValueError('Unknown indexer storage class `%s`' % cls) return IndexerStorage(**args) def check_id_duplicates(data): """ If any two dictionaries in `data` have the same id, raises a `ValueError`. Values associated to the key must be hashable. Args: data (List[dict]): List of dictionaries to be inserted >>> check_id_duplicates([ ... {'id': 'foo', 'data': 'spam'}, ... {'id': 'bar', 'data': 'egg'}, ... ]) >>> check_id_duplicates([ ... {'id': 'foo', 'data': 'spam'}, ... {'id': 'foo', 'data': 'egg'}, ... ]) Traceback (most recent call last): ... swh.indexer.storage.exc.DuplicateId: ['foo'] """ counter = Counter(item['id'] for item in data) duplicates = [id_ for (id_, count) in counter.items() if count >= 2] if duplicates: raise DuplicateId(duplicates) class IndexerStorage: """SWH Indexer Storage """ def __init__(self, db, min_pool_conns=1, max_pool_conns=10): """ Args: db_conn: either a libpq connection string, or a psycopg2 connection """ try: if isinstance(db, psycopg2.extensions.connection): self._pool = None self._db = Db(db) else: self._pool = psycopg2.pool.ThreadedConnectionPool( min_pool_conns, max_pool_conns, db ) self._db = None except psycopg2.OperationalError as e: raise StorageDBError(e) def get_db(self): if self._db: return self._db return Db.from_pool(self._pool) def put_db(self, db): if db is not self._db: db.put_conn() @db_transaction() def check_config(self, *, check_write, db=None, cur=None): # Check permissions on one of the tables if check_write: check = 'INSERT' else: check = 'SELECT' cur.execute( "select has_table_privilege(current_user, 'content_mimetype', %s)", # noqa (check,) ) return cur.fetchone()[0] @db_transaction_generator() def content_mimetype_missing(self, mimetypes, db=None, cur=None): for obj in db.content_mimetype_missing_from_list(mimetypes, cur): yield obj[0] def _content_get_range(self, content_type, start, end, indexer_configuration_id, limit=1000, with_textual_data=False, db=None, cur=None): if limit is None: raise IndexerStorageArgumentException('limit should not be None') if content_type not in db.content_indexer_names: err = 'Wrong type. Should be one of [%s]' % ( ','.join(db.content_indexer_names)) raise IndexerStorageArgumentException(err) ids = [] next_id = None for counter, obj in enumerate(db.content_get_range( content_type, start, end, indexer_configuration_id, limit=limit+1, with_textual_data=with_textual_data, cur=cur)): _id = obj[0] if counter >= limit: next_id = _id break ids.append(_id) return { 'ids': ids, 'next': next_id } @db_transaction() def content_mimetype_get_range(self, start, end, indexer_configuration_id, limit=1000, db=None, cur=None): return self._content_get_range('mimetype', start, end, indexer_configuration_id, limit=limit, db=db, cur=cur) @db_transaction() - def content_mimetype_add(self, mimetypes, conflict_update=False, db=None, - cur=None): + def content_mimetype_add( + self, mimetypes: List[Dict], conflict_update: bool = False, + db=None, cur=None) -> Dict: + """Add mimetypes to the storage (if conflict_update is True, this will + override existing data if any). + + Returns: + A dict with the number of new elements added to the storage. + + """ check_id_duplicates(mimetypes) mimetypes.sort(key=lambda m: m['id']) db.mktemp_content_mimetype(cur) db.copy_to(mimetypes, 'tmp_content_mimetype', ['id', 'mimetype', 'encoding', 'indexer_configuration_id'], cur) - db.content_mimetype_add_from_temp(conflict_update, cur) + count = db.content_mimetype_add_from_temp(conflict_update, cur) + return { + 'content_mimetype:add': count + } @db_transaction_generator() def content_mimetype_get(self, ids, db=None, cur=None): for c in db.content_mimetype_get_from_list(ids, cur): yield converters.db_to_mimetype( dict(zip(db.content_mimetype_cols, c))) @db_transaction_generator() def content_language_missing(self, languages, db=None, cur=None): for obj in db.content_language_missing_from_list(languages, cur): yield obj[0] @db_transaction_generator() def content_language_get(self, ids, db=None, cur=None): for c in db.content_language_get_from_list(ids, cur): yield converters.db_to_language( dict(zip(db.content_language_cols, c))) @db_transaction() - def content_language_add(self, languages, conflict_update=False, db=None, - cur=None): + def content_language_add( + self, languages: List[Dict], + conflict_update: bool = False, db=None, cur=None) -> Dict: check_id_duplicates(languages) languages.sort(key=lambda m: m['id']) db.mktemp_content_language(cur) # empty language is mapped to 'unknown' db.copy_to( ({ 'id': l['id'], 'lang': 'unknown' if not l['lang'] else l['lang'], 'indexer_configuration_id': l['indexer_configuration_id'], } for l in languages), 'tmp_content_language', ['id', 'lang', 'indexer_configuration_id'], cur) - db.content_language_add_from_temp(conflict_update, cur) + count = db.content_language_add_from_temp(conflict_update, cur) + return { + 'content_language:add': count + } @db_transaction_generator() def content_ctags_missing(self, ctags, db=None, cur=None): for obj in db.content_ctags_missing_from_list(ctags, cur): yield obj[0] @db_transaction_generator() def content_ctags_get(self, ids, db=None, cur=None): for c in db.content_ctags_get_from_list(ids, cur): yield converters.db_to_ctags(dict(zip(db.content_ctags_cols, c))) @db_transaction() - def content_ctags_add(self, ctags, conflict_update=False, db=None, - cur=None): + def content_ctags_add( + self, ctags: List[Dict], conflict_update: bool = False, + db=None, cur=None) -> Dict: check_id_duplicates(ctags) ctags.sort(key=lambda m: m['id']) def _convert_ctags(__ctags): """Convert ctags dict to list of ctags. """ for ctags in __ctags: yield from converters.ctags_to_db(ctags) db.mktemp_content_ctags(cur) db.copy_to(list(_convert_ctags(ctags)), tblname='tmp_content_ctags', columns=['id', 'name', 'kind', 'line', 'lang', 'indexer_configuration_id'], cur=cur) - db.content_ctags_add_from_temp(conflict_update, cur) + count = db.content_ctags_add_from_temp(conflict_update, cur) + return { + 'content_ctags:add': count + } @db_transaction_generator() def content_ctags_search(self, expression, limit=10, last_sha1=None, db=None, cur=None): for obj in db.content_ctags_search(expression, last_sha1, limit, cur=cur): yield converters.db_to_ctags(dict(zip(db.content_ctags_cols, obj))) @db_transaction_generator() def content_fossology_license_get(self, ids, db=None, cur=None): d = defaultdict(list) for c in db.content_fossology_license_get_from_list(ids, cur): license = dict(zip(db.content_fossology_license_cols, c)) id_ = license['id'] d[id_].append(converters.db_to_fossology_license(license)) for id_, facts in d.items(): yield {id_: facts} @db_transaction() - def content_fossology_license_add(self, licenses, conflict_update=False, - db=None, cur=None): + def content_fossology_license_add( + self, licenses: List[Dict], conflict_update: bool = False, + db=None, cur=None) -> Dict: check_id_duplicates(licenses) licenses.sort(key=lambda m: m['id']) db.mktemp_content_fossology_license(cur) db.copy_to( ({ 'id': sha1['id'], 'indexer_configuration_id': sha1['indexer_configuration_id'], 'license': license, } for sha1 in licenses for license in sha1['licenses']), tblname='tmp_content_fossology_license', columns=['id', 'license', 'indexer_configuration_id'], cur=cur) - db.content_fossology_license_add_from_temp(conflict_update, cur) + count = db.content_fossology_license_add_from_temp( + conflict_update, cur) + return { + 'content_fossology_license:add': count + } @db_transaction() def content_fossology_license_get_range( self, start, end, indexer_configuration_id, limit=1000, db=None, cur=None): return self._content_get_range('fossology_license', start, end, indexer_configuration_id, limit=limit, with_textual_data=True, db=db, cur=cur) @db_transaction_generator() def content_metadata_missing(self, metadata, db=None, cur=None): for obj in db.content_metadata_missing_from_list(metadata, cur): yield obj[0] @db_transaction_generator() def content_metadata_get(self, ids, db=None, cur=None): for c in db.content_metadata_get_from_list(ids, cur): yield converters.db_to_metadata( dict(zip(db.content_metadata_cols, c))) @db_transaction() - def content_metadata_add(self, metadata, conflict_update=False, db=None, - cur=None): + def content_metadata_add( + self, metadata: List[Dict], conflict_update: bool = False, + db=None, cur=None) -> Dict: check_id_duplicates(metadata) metadata.sort(key=lambda m: m['id']) db.mktemp_content_metadata(cur) db.copy_to(metadata, 'tmp_content_metadata', ['id', 'metadata', 'indexer_configuration_id'], cur) - db.content_metadata_add_from_temp(conflict_update, cur) + count = db.content_metadata_add_from_temp(conflict_update, cur) + return { + 'content_metadata:add': count, + } @db_transaction_generator() def revision_intrinsic_metadata_missing(self, metadata, db=None, cur=None): for obj in db.revision_intrinsic_metadata_missing_from_list( metadata, cur): yield obj[0] @db_transaction_generator() def revision_intrinsic_metadata_get(self, ids, db=None, cur=None): for c in db.revision_intrinsic_metadata_get_from_list(ids, cur): yield converters.db_to_metadata( dict(zip(db.revision_intrinsic_metadata_cols, c))) @db_transaction() - def revision_intrinsic_metadata_add(self, metadata, conflict_update=False, - db=None, cur=None): + def revision_intrinsic_metadata_add( + self, metadata: List[Dict], conflict_update: bool = False, + db=None, cur=None) -> Dict: check_id_duplicates(metadata) metadata.sort(key=lambda m: m['id']) db.mktemp_revision_intrinsic_metadata(cur) db.copy_to(metadata, 'tmp_revision_intrinsic_metadata', ['id', 'metadata', 'mappings', 'indexer_configuration_id'], cur) - db.revision_intrinsic_metadata_add_from_temp(conflict_update, cur) + count = db.revision_intrinsic_metadata_add_from_temp( + conflict_update, cur) + return { + 'revision_intrinsic_metadata:add': count, + } @db_transaction() - def revision_intrinsic_metadata_delete(self, entries, db=None, cur=None): - db.revision_intrinsic_metadata_delete(entries, cur) + def revision_intrinsic_metadata_delete( + self, entries: List[Dict], db=None, cur=None) -> Dict: + count = db.revision_intrinsic_metadata_delete(entries, cur) + return { + 'revision_intrinsic_metadata:del': count + } @db_transaction_generator() def origin_intrinsic_metadata_get(self, ids, db=None, cur=None): for c in db.origin_intrinsic_metadata_get_from_list(ids, cur): yield converters.db_to_metadata( dict(zip(db.origin_intrinsic_metadata_cols, c))) @db_transaction() - def origin_intrinsic_metadata_add(self, metadata, - conflict_update=False, db=None, - cur=None): + def origin_intrinsic_metadata_add( + self, metadata: List[Dict], conflict_update: bool = False, + db=None, cur=None) -> Dict: check_id_duplicates(metadata) metadata.sort(key=lambda m: m['id']) db.mktemp_origin_intrinsic_metadata(cur) db.copy_to(metadata, 'tmp_origin_intrinsic_metadata', ['id', 'metadata', 'indexer_configuration_id', 'from_revision', 'mappings'], cur) - db.origin_intrinsic_metadata_add_from_temp(conflict_update, cur) + count = db.origin_intrinsic_metadata_add_from_temp( + conflict_update, cur) + return { + 'origin_intrinsic_metadata:add': count, + } @db_transaction() def origin_intrinsic_metadata_delete( - self, entries, db=None, cur=None): - db.origin_intrinsic_metadata_delete(entries, cur) + self, entries: List[Dict], db=None, cur=None) -> Dict: + count = db.origin_intrinsic_metadata_delete(entries, cur) + return { + 'origin_intrinsic_metadata:del': count, + } @db_transaction_generator() def origin_intrinsic_metadata_search_fulltext( self, conjunction, limit=100, db=None, cur=None): for c in db.origin_intrinsic_metadata_search_fulltext( conjunction, limit=limit, cur=cur): yield converters.db_to_metadata( dict(zip(db.origin_intrinsic_metadata_cols, c))) @db_transaction() def origin_intrinsic_metadata_search_by_producer( self, page_token='', limit=100, ids_only=False, mappings=None, tool_ids=None, db=None, cur=None): assert isinstance(page_token, str) # we go to limit+1 to check whether we should add next_page_token in # the response res = db.origin_intrinsic_metadata_search_by_producer( page_token, limit + 1, ids_only, mappings, tool_ids, cur) result = {} if ids_only: result['origins'] = [origin for (origin,) in res] if len(result['origins']) > limit: result['origins'][limit:] = [] result['next_page_token'] = result['origins'][-1] else: result['origins'] = [converters.db_to_metadata( dict(zip(db.origin_intrinsic_metadata_cols, c)))for c in res] if len(result['origins']) > limit: result['origins'][limit:] = [] result['next_page_token'] = result['origins'][-1]['id'] return result @db_transaction() def origin_intrinsic_metadata_stats( self, db=None, cur=None): mapping_names = [m for m in MAPPING_NAMES] select_parts = [] # Count rows for each mapping for mapping_name in mapping_names: select_parts.append(( "sum(case when (mappings @> ARRAY['%s']) " " then 1 else 0 end)" ) % mapping_name) # Total select_parts.append("sum(1)") # Rows whose metadata has at least one key that is not '@context' select_parts.append( "sum(case when ('{}'::jsonb @> (metadata - '@context')) " " then 0 else 1 end)") cur.execute('select ' + ', '.join(select_parts) + ' from origin_intrinsic_metadata') results = dict(zip(mapping_names + ['total', 'non_empty'], cur.fetchone())) return { 'total': results.pop('total'), 'non_empty': results.pop('non_empty'), 'per_mapping': results, } @db_transaction_generator() def indexer_configuration_add(self, tools, db=None, cur=None): db.mktemp_indexer_configuration(cur) db.copy_to(tools, 'tmp_indexer_configuration', ['tool_name', 'tool_version', 'tool_configuration'], cur) tools = db.indexer_configuration_add_from_temp(cur) for line in tools: yield dict(zip(db.indexer_configuration_cols, line)) @db_transaction() def indexer_configuration_get(self, tool, db=None, cur=None): tool_conf = tool['tool_configuration'] if isinstance(tool_conf, dict): tool_conf = json.dumps(tool_conf) idx = db.indexer_configuration_get(tool['tool_name'], tool['tool_version'], tool_conf) if not idx: return None return dict(zip(db.indexer_configuration_cols, idx)) diff --git a/swh/indexer/storage/db.py b/swh/indexer/storage/db.py index 241ee08..80e14ef 100644 --- a/swh/indexer/storage/db.py +++ b/swh/indexer/storage/db.py @@ -1,460 +1,479 @@ # Copyright (C) 2015-2018 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information from swh.model import hashutil from swh.core.db import BaseDb from swh.core.db.db_utils import execute_values_generator, stored_procedure class Db(BaseDb): """Proxy to the SWH Indexer DB, with wrappers around stored procedures """ content_mimetype_hash_keys = ['id', 'indexer_configuration_id'] def _missing_from_list(self, table, data, hash_keys, cur=None): """Read from table the data with hash_keys that are missing. Args: table (str): Table name (e.g content_mimetype, content_language, etc...) data (dict): Dict of data to read from hash_keys ([str]): List of keys to read in the data dict. Yields: The data which is missing from the db. """ cur = self._cursor(cur) keys = ', '.join(hash_keys) equality = ' AND '.join( ('t.%s = c.%s' % (key, key)) for key in hash_keys ) yield from execute_values_generator( cur, """ select %s from (values %%s) as t(%s) where not exists ( select 1 from %s c where %s ) """ % (keys, keys, table, equality), (tuple(m[k] for k in hash_keys) for m in data) ) def content_mimetype_missing_from_list(self, mimetypes, cur=None): """List missing mimetypes. """ yield from self._missing_from_list( 'content_mimetype', mimetypes, self.content_mimetype_hash_keys, cur=cur) content_mimetype_cols = [ 'id', 'mimetype', 'encoding', 'tool_id', 'tool_name', 'tool_version', 'tool_configuration'] @stored_procedure('swh_mktemp_content_mimetype') def mktemp_content_mimetype(self, cur=None): pass def content_mimetype_add_from_temp(self, conflict_update, cur=None): - self._cursor(cur).execute("SELECT swh_content_mimetype_add(%s)", - (conflict_update, )) + cur = self._cursor(cur) + cur.execute('select * from swh_content_mimetype_add(%s)', + (conflict_update, )) + return cur.fetchone()[0] def _convert_key(self, key, main_table='c'): """Convert keys according to specific use in the module. Args: key (str): Key expression to change according to the alias used in the query main_table (str): Alias to use for the main table. Default to c for content_{something}. Expected: Tables content_{something} being aliased as 'c' (something in {language, mimetype, ...}), table indexer_configuration being aliased as 'i'. """ if key == 'id': return '%s.id' % main_table elif key == 'tool_id': return 'i.id as tool_id' elif key == 'licenses': return ''' array(select name from fossology_license where id = ANY( array_agg(%s.license_id))) as licenses''' % main_table return key def _get_from_list(self, table, ids, cols, cur=None, id_col='id'): """Fetches entries from the `table` such that their `id` field (or whatever is given to `id_col`) is in `ids`. Returns the columns `cols`. The `cur` parameter is used to connect to the database. """ cur = self._cursor(cur) keys = map(self._convert_key, cols) query = """ select {keys} from (values %s) as t(id) inner join {table} c on c.{id_col}=t.id inner join indexer_configuration i on c.indexer_configuration_id=i.id; """.format( keys=', '.join(keys), id_col=id_col, table=table) yield from execute_values_generator( cur, query, ((_id,) for _id in ids) ) content_indexer_names = { 'mimetype': 'content_mimetype', 'fossology_license': 'content_fossology_license', } def content_get_range(self, content_type, start, end, indexer_configuration_id, limit=1000, with_textual_data=False, cur=None): """Retrieve contents with content_type, within range [start, end] bound by limit and associated to the given indexer configuration id. When asking to work on textual content, that filters on the mimetype table with any mimetype that is not binary. """ cur = self._cursor(cur) table = self.content_indexer_names[content_type] if with_textual_data: extra = """inner join content_mimetype cm on (t.id=cm.id and cm.mimetype like 'text/%%' and %(start)s <= cm.id and cm.id <= %(end)s) """ else: extra = "" query = f"""select t.id from {table} t {extra} where t.indexer_configuration_id=%(tool_id)s and %(start)s <= t.id and t.id <= %(end)s order by t.indexer_configuration_id, t.id limit %(limit)s""" cur.execute(query, { 'start': start, 'end': end, 'tool_id': indexer_configuration_id, 'limit': limit, }) yield from cur def content_mimetype_get_from_list(self, ids, cur=None): yield from self._get_from_list( 'content_mimetype', ids, self.content_mimetype_cols, cur=cur) content_language_hash_keys = ['id', 'indexer_configuration_id'] def content_language_missing_from_list(self, languages, cur=None): """List missing languages. """ yield from self._missing_from_list( 'content_language', languages, self.content_language_hash_keys, cur=cur) content_language_cols = [ 'id', 'lang', 'tool_id', 'tool_name', 'tool_version', 'tool_configuration'] @stored_procedure('swh_mktemp_content_language') def mktemp_content_language(self, cur=None): pass def content_language_add_from_temp(self, conflict_update, cur=None): - self._cursor(cur).execute("SELECT swh_content_language_add(%s)", - (conflict_update, )) + cur = self._cursor(cur) + cur.execute('select * from swh_content_language_add(%s)', + (conflict_update, )) + return cur.fetchone()[0] def content_language_get_from_list(self, ids, cur=None): yield from self._get_from_list( 'content_language', ids, self.content_language_cols, cur=cur) content_ctags_hash_keys = ['id', 'indexer_configuration_id'] def content_ctags_missing_from_list(self, ctags, cur=None): """List missing ctags. """ yield from self._missing_from_list( 'content_ctags', ctags, self.content_ctags_hash_keys, cur=cur) content_ctags_cols = [ 'id', 'name', 'kind', 'line', 'lang', 'tool_id', 'tool_name', 'tool_version', 'tool_configuration'] @stored_procedure('swh_mktemp_content_ctags') def mktemp_content_ctags(self, cur=None): pass def content_ctags_add_from_temp(self, conflict_update, cur=None): - self._cursor(cur).execute("SELECT swh_content_ctags_add(%s)", - (conflict_update, )) + cur = self._cursor(cur) + cur.execute( + 'select * from swh_content_ctags_add(%s)', + (conflict_update, )) + return cur.fetchone()[0] def content_ctags_get_from_list(self, ids, cur=None): cur = self._cursor(cur) keys = map(self._convert_key, self.content_ctags_cols) yield from execute_values_generator( cur, """ select %s from (values %%s) as t(id) inner join content_ctags c on c.id=t.id inner join indexer_configuration i on c.indexer_configuration_id=i.id order by line """ % ', '.join(keys), ((_id,) for _id in ids) ) def content_ctags_search(self, expression, last_sha1, limit, cur=None): cur = self._cursor(cur) if not last_sha1: query = """SELECT %s FROM swh_content_ctags_search(%%s, %%s)""" % ( ','.join(self.content_ctags_cols)) cur.execute(query, (expression, limit)) else: if last_sha1 and isinstance(last_sha1, bytes): last_sha1 = '\\x%s' % hashutil.hash_to_hex(last_sha1) elif last_sha1: last_sha1 = '\\x%s' % last_sha1 query = """SELECT %s FROM swh_content_ctags_search(%%s, %%s, %%s)""" % ( ','.join(self.content_ctags_cols)) cur.execute(query, (expression, limit, last_sha1)) yield from cur content_fossology_license_cols = [ 'id', 'tool_id', 'tool_name', 'tool_version', 'tool_configuration', 'licenses'] @stored_procedure('swh_mktemp_content_fossology_license') def mktemp_content_fossology_license(self, cur=None): pass def content_fossology_license_add_from_temp(self, conflict_update, cur=None): """Add new licenses per content. """ - self._cursor(cur).execute( - "SELECT swh_content_fossology_license_add(%s)", + cur = self._cursor(cur) + cur.execute( + 'select * from swh_content_fossology_license_add(%s)', (conflict_update, )) + return cur.fetchone()[0] def content_fossology_license_get_from_list(self, ids, cur=None): """Retrieve licenses per id. """ cur = self._cursor(cur) keys = map(self._convert_key, self.content_fossology_license_cols) yield from execute_values_generator( cur, """ select %s from (values %%s) as t(id) inner join content_fossology_license c on t.id=c.id inner join indexer_configuration i on i.id=c.indexer_configuration_id group by c.id, i.id, i.tool_name, i.tool_version, i.tool_configuration; """ % ', '.join(keys), ((_id,) for _id in ids) ) content_metadata_hash_keys = ['id', 'indexer_configuration_id'] def content_metadata_missing_from_list(self, metadata, cur=None): """List missing metadata. """ yield from self._missing_from_list( 'content_metadata', metadata, self.content_metadata_hash_keys, cur=cur) content_metadata_cols = [ 'id', 'metadata', 'tool_id', 'tool_name', 'tool_version', 'tool_configuration'] @stored_procedure('swh_mktemp_content_metadata') def mktemp_content_metadata(self, cur=None): pass def content_metadata_add_from_temp(self, conflict_update, cur=None): - self._cursor(cur).execute("SELECT swh_content_metadata_add(%s)", - (conflict_update, )) + cur = self._cursor(cur) + cur.execute( + 'select * from swh_content_metadata_add(%s)', + (conflict_update, )) + return cur.fetchone()[0] def content_metadata_get_from_list(self, ids, cur=None): yield from self._get_from_list( 'content_metadata', ids, self.content_metadata_cols, cur=cur) revision_intrinsic_metadata_hash_keys = [ 'id', 'indexer_configuration_id'] def revision_intrinsic_metadata_missing_from_list( self, metadata, cur=None): """List missing metadata. """ yield from self._missing_from_list( 'revision_intrinsic_metadata', metadata, self.revision_intrinsic_metadata_hash_keys, cur=cur) revision_intrinsic_metadata_cols = [ 'id', 'metadata', 'mappings', 'tool_id', 'tool_name', 'tool_version', 'tool_configuration'] @stored_procedure('swh_mktemp_revision_intrinsic_metadata') def mktemp_revision_intrinsic_metadata(self, cur=None): pass def revision_intrinsic_metadata_add_from_temp( self, conflict_update, cur=None): - self._cursor(cur).execute( - "SELECT swh_revision_intrinsic_metadata_add(%s)", - (conflict_update, )) + cur = self._cursor(cur) + cur.execute( + 'select * from swh_revision_intrinsic_metadata_add(%s)', + (conflict_update, )) + return cur.fetchone()[0] def revision_intrinsic_metadata_delete( self, entries, cur=None): cur = self._cursor(cur) cur.execute( "DELETE from revision_intrinsic_metadata " "WHERE (id, indexer_configuration_id) IN " - " (VALUES %s)" % (', '.join('%s' for _ in entries)), + " (VALUES %s) " + "RETURNING id" % (', '.join('%s' for _ in entries)), tuple((e['id'], e['indexer_configuration_id']) for e in entries),) + return len(cur.fetchall()) def revision_intrinsic_metadata_get_from_list(self, ids, cur=None): yield from self._get_from_list( 'revision_intrinsic_metadata', ids, self.revision_intrinsic_metadata_cols, cur=cur) origin_intrinsic_metadata_cols = [ 'id', 'metadata', 'from_revision', 'mappings', 'tool_id', 'tool_name', 'tool_version', 'tool_configuration'] origin_intrinsic_metadata_regconfig = 'pg_catalog.simple' """The dictionary used to normalize 'metadata' and queries. 'pg_catalog.simple' provides no stopword, so it should be suitable for proper names and non-English content. When updating this value, make sure to add a new index on origin_intrinsic_metadata.metadata.""" @stored_procedure('swh_mktemp_origin_intrinsic_metadata') def mktemp_origin_intrinsic_metadata(self, cur=None): pass def origin_intrinsic_metadata_add_from_temp( self, conflict_update, cur=None): cur = self._cursor(cur) cur.execute( - "SELECT swh_origin_intrinsic_metadata_add(%s)", - (conflict_update, )) + 'select * from swh_origin_intrinsic_metadata_add(%s)', + (conflict_update, )) + return cur.fetchone()[0] def origin_intrinsic_metadata_delete( self, entries, cur=None): cur = self._cursor(cur) cur.execute( "DELETE from origin_intrinsic_metadata " "WHERE (id, indexer_configuration_id) IN" - " (VALUES %s)" % (', '.join('%s' for _ in entries)), + " (VALUES %s) " + "RETURNING id" % (', '.join('%s' for _ in entries)), tuple((e['id'], e['indexer_configuration_id']) for e in entries),) + return len(cur.fetchall()) def origin_intrinsic_metadata_get_from_list(self, ids, cur=None): yield from self._get_from_list( 'origin_intrinsic_metadata', ids, self.origin_intrinsic_metadata_cols, cur=cur, id_col='id') def origin_intrinsic_metadata_search_fulltext(self, terms, *, limit, cur): regconfig = self.origin_intrinsic_metadata_regconfig tsquery_template = ' && '.join("plainto_tsquery('%s', %%s)" % regconfig for _ in terms) tsquery_args = [(term,) for term in terms] keys = (self._convert_key(col, 'oim') for col in self.origin_intrinsic_metadata_cols) query = ("SELECT {keys} FROM origin_intrinsic_metadata AS oim " "INNER JOIN indexer_configuration AS i " "ON oim.indexer_configuration_id=i.id " "JOIN LATERAL (SELECT {tsquery_template}) AS s(tsq) ON true " "WHERE oim.metadata_tsvector @@ tsq " "ORDER BY ts_rank(oim.metadata_tsvector, tsq, 1) DESC " "LIMIT %s;" ).format(keys=', '.join(keys), tsquery_template=tsquery_template) cur.execute(query, tsquery_args + [limit]) yield from cur def origin_intrinsic_metadata_search_by_producer( self, last, limit, ids_only, mappings, tool_ids, cur): if ids_only: keys = 'oim.id' else: keys = ', '.join((self._convert_key(col, 'oim') for col in self.origin_intrinsic_metadata_cols)) query_parts = [ "SELECT %s" % keys, "FROM origin_intrinsic_metadata AS oim", "INNER JOIN indexer_configuration AS i", "ON oim.indexer_configuration_id=i.id", ] args = [] where = [] if last: where.append('oim.id > %s') args.append(last) if mappings is not None: where.append('oim.mappings && %s') args.append(mappings) if tool_ids is not None: where.append('oim.indexer_configuration_id = ANY(%s)') args.append(tool_ids) if where: query_parts.append('WHERE') query_parts.append(' AND '.join(where)) if limit: query_parts.append('LIMIT %s') args.append(limit) cur.execute(' '.join(query_parts), args) yield from cur indexer_configuration_cols = ['id', 'tool_name', 'tool_version', 'tool_configuration'] @stored_procedure('swh_mktemp_indexer_configuration') def mktemp_indexer_configuration(self, cur=None): pass def indexer_configuration_add_from_temp(self, cur=None): cur = self._cursor(cur) cur.execute("SELECT %s from swh_indexer_configuration_add()" % ( ','.join(self.indexer_configuration_cols), )) yield from cur def indexer_configuration_get(self, tool_name, tool_version, tool_configuration, cur=None): cur = self._cursor(cur) cur.execute('''select %s from indexer_configuration where tool_name=%%s and tool_version=%%s and tool_configuration=%%s''' % ( ','.join(self.indexer_configuration_cols)), (tool_name, tool_version, tool_configuration)) return cur.fetchone() diff --git a/swh/indexer/storage/in_memory.py b/swh/indexer/storage/in_memory.py index 46e1516..225b53e 100644 --- a/swh/indexer/storage/in_memory.py +++ b/swh/indexer/storage/in_memory.py @@ -1,421 +1,451 @@ -# Copyright (C) 2018 The Software Heritage developers +# Copyright (C) 2018-2020 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import bisect from collections import defaultdict, Counter import itertools import json import operator import math import re from typing import Any, Dict, List from . import MAPPING_NAMES, check_id_duplicates from .exc import IndexerStorageArgumentException SHA1_DIGEST_SIZE = 160 def _transform_tool(tool): return { 'id': tool['id'], 'name': tool['tool_name'], 'version': tool['tool_version'], 'configuration': tool['tool_configuration'], } def check_id_types(data: List[Dict[str, Any]]): """Checks all elements of the list have an 'id' whose type is 'bytes'.""" if not all(isinstance(item.get('id'), bytes) for item in data): raise IndexerStorageArgumentException('identifiers must be bytes.') class SubStorage: """Implements common missing/get/add logic for each indexer type.""" def __init__(self, tools): self._tools = tools self._sorted_ids = [] self._data = {} # map (id_, tool_id) -> metadata_dict self._tools_per_id = defaultdict(set) # map id_ -> Set[tool_id] def missing(self, ids): """List data missing from storage. Args: data (iterable): dictionaries with keys: - **id** (bytes): sha1 identifier - **indexer_configuration_id** (int): tool used to compute the results Yields: missing sha1s """ for id_ in ids: tool_id = id_['indexer_configuration_id'] id_ = id_['id'] if tool_id not in self._tools_per_id.get(id_, set()): yield id_ def get(self, ids): """Retrieve data per id. Args: ids (iterable): sha1 checksums Yields: dict: dictionaries with the following keys: - **id** (bytes) - **tool** (dict): tool used to compute metadata - arbitrary data (as provided to `add`) """ for id_ in ids: for tool_id in self._tools_per_id.get(id_, set()): key = (id_, tool_id) yield { 'id': id_, 'tool': _transform_tool(self._tools[tool_id]), **self._data[key], } def get_all(self): yield from self.get(self._sorted_ids) def get_range(self, start, end, indexer_configuration_id, limit): """Retrieve data within range [start, end] bound by limit. Args: **start** (bytes): Starting identifier range (expected smaller than end) **end** (bytes): Ending identifier range (expected larger than start) **indexer_configuration_id** (int): The tool used to index data **limit** (int): Limit result Raises: IndexerStorageArgumentException for limit to None Returns: a dict with keys: - **ids** [bytes]: iterable of content ids within the range. - **next** (Optional[bytes]): The next range of sha1 starts at this sha1 if any """ if limit is None: raise IndexerStorageArgumentException('limit should not be None') from_index = bisect.bisect_left(self._sorted_ids, start) to_index = bisect.bisect_right(self._sorted_ids, end, lo=from_index) if to_index - from_index >= limit: return { 'ids': self._sorted_ids[from_index:from_index+limit], 'next': self._sorted_ids[from_index+limit], } else: return { 'ids': self._sorted_ids[from_index:to_index], 'next': None, } - def add(self, data, conflict_update): + def add(self, data: List[Dict], conflict_update: bool) -> int: """Add data not present in storage. Args: data (iterable): dictionaries with keys: - **id**: sha1 - **indexer_configuration_id**: tool used to compute the results - arbitrary data conflict_update (bool): Flag to determine if we want to overwrite (true) or skip duplicates (false) """ data = list(data) check_id_duplicates(data) + count = 0 for item in data: item = item.copy() tool_id = item.pop('indexer_configuration_id') id_ = item.pop('id') - data = item + data_item = item if not conflict_update and \ tool_id in self._tools_per_id.get(id_, set()): # Duplicate, should not be updated continue key = (id_, tool_id) - self._data[key] = data + self._data[key] = data_item self._tools_per_id[id_].add(tool_id) + count += 1 if id_ not in self._sorted_ids: bisect.insort(self._sorted_ids, id_) + return count - def add_merge(self, new_data, conflict_update, merged_key): + def add_merge(self, new_data: List[Dict], conflict_update: bool, + merged_key: str) -> int: + added = 0 for new_item in new_data: id_ = new_item['id'] tool_id = new_item['indexer_configuration_id'] if conflict_update: all_subitems = [] else: existing = list(self.get([id_])) all_subitems = [ old_subitem for existing_item in existing if existing_item['tool']['id'] == tool_id for old_subitem in existing_item[merged_key] ] for new_subitem in new_item[merged_key]: if new_subitem not in all_subitems: all_subitems.append(new_subitem) - self.add([ + added += self.add([ { 'id': id_, 'indexer_configuration_id': tool_id, merged_key: all_subitems, } ], conflict_update=True) if id_ not in self._sorted_ids: bisect.insort(self._sorted_ids, id_) + return added - def delete(self, entries): + def delete(self, entries: List[Dict]) -> int: + """Delete entries and return the number of entries deleted. + + """ + deleted = 0 for entry in entries: (id_, tool_id) = (entry['id'], entry['indexer_configuration_id']) key = (id_, tool_id) if tool_id in self._tools_per_id[id_]: self._tools_per_id[id_].remove(tool_id) if key in self._data: + deleted += 1 del self._data[key] + return deleted class IndexerStorage: """In-memory SWH indexer storage.""" def __init__(self): self._tools = {} self._mimetypes = SubStorage(self._tools) self._languages = SubStorage(self._tools) self._content_ctags = SubStorage(self._tools) self._licenses = SubStorage(self._tools) self._content_metadata = SubStorage(self._tools) self._revision_intrinsic_metadata = SubStorage(self._tools) self._origin_intrinsic_metadata = SubStorage(self._tools) def check_config(self, *, check_write): return True def content_mimetype_missing(self, mimetypes): yield from self._mimetypes.missing(mimetypes) def content_mimetype_get_range( self, start, end, indexer_configuration_id, limit=1000): return self._mimetypes.get_range( start, end, indexer_configuration_id, limit) - def content_mimetype_add(self, mimetypes, conflict_update=False): + def content_mimetype_add( + self, mimetypes: List[Dict], + conflict_update: bool = False) -> Dict: check_id_types(mimetypes) - self._mimetypes.add(mimetypes, conflict_update) + added = self._mimetypes.add(mimetypes, conflict_update) + return {'content_mimetype:add': added} def content_mimetype_get(self, ids): yield from self._mimetypes.get(ids) def content_language_missing(self, languages): yield from self._languages.missing(languages) def content_language_get(self, ids): yield from self._languages.get(ids) - def content_language_add(self, languages, conflict_update=False): + def content_language_add( + self, languages: List[Dict], + conflict_update: bool = False) -> Dict: check_id_types(languages) - self._languages.add(languages, conflict_update) + added = self._languages.add(languages, conflict_update) + return {'content_language:add': added} def content_ctags_missing(self, ctags): yield from self._content_ctags.missing(ctags) def content_ctags_get(self, ids): for item in self._content_ctags.get(ids): for item_ctags_item in item['ctags']: yield { 'id': item['id'], 'tool': item['tool'], **item_ctags_item } - def content_ctags_add(self, ctags, conflict_update=False): + def content_ctags_add( + self, ctags: List[Dict], conflict_update: bool = False) -> Dict: check_id_types(ctags) - self._content_ctags.add_merge(ctags, conflict_update, 'ctags') + added = self._content_ctags.add_merge(ctags, conflict_update, 'ctags') + return {'content_ctags:add': added} def content_ctags_search(self, expression, limit=10, last_sha1=None): nb_matches = 0 for ((id_, tool_id), item) in \ sorted(self._content_ctags._data.items()): if id_ <= (last_sha1 or bytes(0 for _ in range(SHA1_DIGEST_SIZE))): continue for ctags_item in item['ctags']: if ctags_item['name'] != expression: continue nb_matches += 1 yield { 'id': id_, 'tool': _transform_tool(self._tools[tool_id]), **ctags_item } if nb_matches >= limit: return def content_fossology_license_get(self, ids): # Rewrites the output of SubStorage.get from the old format to # the new one. SubStorage.get should be updated once all other # *_get methods use the new format. # See: https://forge.softwareheritage.org/T1433 res = {} for d in self._licenses.get(ids): res.setdefault(d.pop('id'), []).append(d) for (id_, facts) in res.items(): yield {id_: facts} - def content_fossology_license_add(self, licenses, conflict_update=False): + def content_fossology_license_add( + self, licenses: List[Dict], conflict_update: bool = False) -> Dict: check_id_types(licenses) - self._licenses.add_merge(licenses, conflict_update, 'licenses') + added = self._licenses.add_merge(licenses, conflict_update, 'licenses') + return {'fossology_license_add:add': added} def content_fossology_license_get_range( self, start, end, indexer_configuration_id, limit=1000): return self._licenses.get_range( start, end, indexer_configuration_id, limit) def content_metadata_missing(self, metadata): yield from self._content_metadata.missing(metadata) def content_metadata_get(self, ids): yield from self._content_metadata.get(ids) - def content_metadata_add(self, metadata, conflict_update=False): + def content_metadata_add( + self, metadata: List[Dict], conflict_update: bool = False) -> Dict: check_id_types(metadata) - self._content_metadata.add(metadata, conflict_update) + added = self._content_metadata.add(metadata, conflict_update) + return {'content_metadata:add': added} def revision_intrinsic_metadata_missing(self, metadata): yield from self._revision_intrinsic_metadata.missing(metadata) def revision_intrinsic_metadata_get(self, ids): yield from self._revision_intrinsic_metadata.get(ids) - def revision_intrinsic_metadata_add(self, metadata, conflict_update=False): + def revision_intrinsic_metadata_add( + self, metadata: List[Dict], conflict_update: bool = False) -> Dict: check_id_types(metadata) - self._revision_intrinsic_metadata.add(metadata, conflict_update) + added = self._revision_intrinsic_metadata.add( + metadata, conflict_update) + return {'revision_intrinsic_metadata:add': added} - def revision_intrinsic_metadata_delete(self, entries): - self._revision_intrinsic_metadata.delete(entries) + def revision_intrinsic_metadata_delete(self, entries: List[Dict]) -> Dict: + deleted = self._revision_intrinsic_metadata.delete(entries) + return {'revision_intrinsic_metadata:del': deleted} def origin_intrinsic_metadata_get(self, ids): yield from self._origin_intrinsic_metadata.get(ids) - def origin_intrinsic_metadata_add(self, metadata, - conflict_update=False): - self._origin_intrinsic_metadata.add(metadata, conflict_update) + def origin_intrinsic_metadata_add( + self, metadata: List[Dict], conflict_update: bool = False) -> Dict: + added = self._origin_intrinsic_metadata.add(metadata, conflict_update) + return {'origin_intrinsic_metadata:add': added} - def origin_intrinsic_metadata_delete(self, entries): - self._origin_intrinsic_metadata.delete(entries) + def origin_intrinsic_metadata_delete(self, entries: List[Dict]) -> Dict: + deleted = self._origin_intrinsic_metadata.delete(entries) + return {'origin_intrinsic_metadata:del': deleted} def origin_intrinsic_metadata_search_fulltext( self, conjunction, limit=100): # A very crude fulltext search implementation, but that's enough # to work on English metadata tokens_re = re.compile('[a-zA-Z0-9]+') search_tokens = list(itertools.chain( *map(tokens_re.findall, conjunction))) def rank(data): # Tokenize the metadata text = json.dumps(data['metadata']) text_tokens = tokens_re.findall(text) text_token_occurences = Counter(text_tokens) # Count the number of occurrences of search tokens in the text score = 0 for search_token in search_tokens: if text_token_occurences[search_token] == 0: # Search token is not in the text. return 0 score += text_token_occurences[search_token] # Normalize according to the text's length return score / math.log(len(text_tokens)) results = [(rank(data), data) for data in self._origin_intrinsic_metadata.get_all()] results = [(rank_, data) for (rank_, data) in results if rank_ > 0] results.sort(key=operator.itemgetter(0), # Don't try to order 'data' reverse=True) for (rank_, result) in results[:limit]: yield result def origin_intrinsic_metadata_search_by_producer( self, page_token='', limit=100, ids_only=False, mappings=None, tool_ids=None): assert isinstance(page_token, str) nb_results = 0 if mappings is not None: mappings = frozenset(mappings) if tool_ids is not None: tool_ids = frozenset(tool_ids) origins = [] # we go to limit+1 to check whether we should add next_page_token in # the response for entry in self._origin_intrinsic_metadata.get_all(): if entry['id'] <= page_token: continue if nb_results >= (limit + 1): break if mappings is not None and mappings.isdisjoint(entry['mappings']): continue if tool_ids is not None and entry['tool']['id'] not in tool_ids: continue origins.append(entry) nb_results += 1 result = {} if len(origins) > limit: origins = origins[:limit] result['next_page_token'] = origins[-1]['id'] if ids_only: origins = [origin['id'] for origin in origins] result['origins'] = origins return result def origin_intrinsic_metadata_stats(self): mapping_count = {m: 0 for m in MAPPING_NAMES} total = non_empty = 0 for data in self._origin_intrinsic_metadata.get_all(): total += 1 if set(data['metadata']) - {'@context'}: non_empty += 1 for mapping in data['mappings']: mapping_count[mapping] += 1 return { 'per_mapping': mapping_count, 'total': total, 'non_empty': non_empty } def indexer_configuration_add(self, tools): inserted = [] for tool in tools: tool = tool.copy() id_ = self._tool_key(tool) tool['id'] = id_ self._tools[id_] = tool inserted.append(tool) return inserted def indexer_configuration_get(self, tool): return self._tools.get(self._tool_key(tool)) def _tool_key(self, tool): return hash((tool['tool_name'], tool['tool_version'], json.dumps(tool['tool_configuration'], sort_keys=True))) diff --git a/swh/indexer/storage/interface.py b/swh/indexer/storage/interface.py index a884f1c..49c24c7 100644 --- a/swh/indexer/storage/interface.py +++ b/swh/indexer/storage/interface.py @@ -1,600 +1,632 @@ # Copyright (C) 2015-2020 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information +from typing import Dict, List from swh.core.api import remote_api_endpoint class IndexerStorageInterface: @remote_api_endpoint('check_config') def check_config(self, *, check_write): """Check that the storage is configured and ready to go.""" ... @remote_api_endpoint('content_mimetype/missing') def content_mimetype_missing(self, mimetypes): """Generate mimetypes missing from storage. Args: mimetypes (iterable): iterable of dict with keys: - **id** (bytes): sha1 identifier - **indexer_configuration_id** (int): tool used to compute the results Yields: tuple (id, indexer_configuration_id): missing id """ ... def _content_get_range(self, content_type, start, end, indexer_configuration_id, limit=1000, with_textual_data=False): """Retrieve ids of type content_type within range [start, end] bound by limit. Args: **content_type** (str): content's type (mimetype, language, etc...) **start** (bytes): Starting identifier range (expected smaller than end) **end** (bytes): Ending identifier range (expected larger than start) **indexer_configuration_id** (int): The tool used to index data **limit** (int): Limit result (default to 1000) **with_textual_data** (bool): Deal with only textual content (True) or all content (all contents by defaults, False) Raises: ValueError for; - limit to None - wrong content_type provided Returns: a dict with keys: - **ids** [bytes]: iterable of content ids within the range. - **next** (Optional[bytes]): The next range of sha1 starts at this sha1 if any """ ... @remote_api_endpoint('content_mimetype/range') def content_mimetype_get_range(self, start, end, indexer_configuration_id, limit=1000): """Retrieve mimetypes within range [start, end] bound by limit. Args: **start** (bytes): Starting identifier range (expected smaller than end) **end** (bytes): Ending identifier range (expected larger than start) **indexer_configuration_id** (int): The tool used to index data **limit** (int): Limit result (default to 1000) Raises: ValueError for limit to None Returns: a dict with keys: - **ids** [bytes]: iterable of content ids within the range. - **next** (Optional[bytes]): The next range of sha1 starts at this sha1 if any """ ... @remote_api_endpoint('content_mimetype/add') - def content_mimetype_add(self, mimetypes, conflict_update=False): + def content_mimetype_add(self, mimetypes: List[Dict], + conflict_update: bool = False) -> Dict: """Add mimetypes not present in storage. Args: mimetypes (iterable): dictionaries with keys: - **id** (bytes): sha1 identifier - **mimetype** (bytes): raw content's mimetype - **encoding** (bytes): raw content's encoding - **indexer_configuration_id** (int): tool's id used to compute the results - **conflict_update** (bool): Flag to determine if we want to overwrite (``True``) or skip duplicates (``False``, the default) + Returns: + Dict summary of number of rows added + """ ... @remote_api_endpoint('content_mimetype') def content_mimetype_get(self, ids): """Retrieve full content mimetype per ids. Args: ids (iterable): sha1 identifier Yields: mimetypes (iterable): dictionaries with keys: - **id** (bytes): sha1 identifier - **mimetype** (bytes): raw content's mimetype - **encoding** (bytes): raw content's encoding - **tool** (dict): Tool used to compute the language """ ... @remote_api_endpoint('content_language/missing') def content_language_missing(self, languages): """List languages missing from storage. Args: languages (iterable): dictionaries with keys: - **id** (bytes): sha1 identifier - **indexer_configuration_id** (int): tool used to compute the results Yields: an iterable of missing id for the tuple (id, indexer_configuration_id) """ ... @remote_api_endpoint('content_language') def content_language_get(self, ids): """Retrieve full content language per ids. Args: ids (iterable): sha1 identifier Yields: languages (iterable): dictionaries with keys: - **id** (bytes): sha1 identifier - **lang** (bytes): raw content's language - **tool** (dict): Tool used to compute the language """ ... @remote_api_endpoint('content_language/add') - def content_language_add(self, languages, conflict_update=False): + def content_language_add( + self, languages: List[Dict], + conflict_update: bool = False) -> Dict: """Add languages not present in storage. Args: languages (iterable): dictionaries with keys: - **id** (bytes): sha1 - **lang** (bytes): language detected conflict_update (bool): Flag to determine if we want to overwrite (true) or skip duplicates (false, the default) + Returns: + Dict summary of number of rows added + """ ... @remote_api_endpoint('content/ctags/missing') def content_ctags_missing(self, ctags): """List ctags missing from storage. Args: ctags (iterable): dicts with keys: - **id** (bytes): sha1 identifier - **indexer_configuration_id** (int): tool used to compute the results Yields: an iterable of missing id for the tuple (id, indexer_configuration_id) """ ... @remote_api_endpoint('content/ctags') def content_ctags_get(self, ids): """Retrieve ctags per id. Args: ids (iterable): sha1 checksums Yields: Dictionaries with keys: - **id** (bytes): content's identifier - **name** (str): symbol's name - **kind** (str): symbol's kind - **lang** (str): language for that content - **tool** (dict): tool used to compute the ctags' info """ ... @remote_api_endpoint('content/ctags/add') - def content_ctags_add(self, ctags, conflict_update=False): + def content_ctags_add(self, ctags: List[Dict], + conflict_update: bool = False) -> Dict: """Add ctags not present in storage Args: ctags (iterable): dictionaries with keys: - **id** (bytes): sha1 - **ctags** ([list): List of dictionary with keys: name, kind, line, lang + Returns: + Dict summary of number of rows added + """ ... @remote_api_endpoint('content/ctags/search') def content_ctags_search(self, expression, limit=10, last_sha1=None): """Search through content's raw ctags symbols. Args: expression (str): Expression to search for limit (int): Number of rows to return (default to 10). last_sha1 (str): Offset from which retrieving data (default to ''). Yields: rows of ctags including id, name, lang, kind, line, etc... """ ... @remote_api_endpoint('content/fossology_license') def content_fossology_license_get(self, ids): """Retrieve licenses per id. Args: ids (iterable): sha1 checksums Yields: dict: ``{id: facts}`` where ``facts`` is a dict with the following keys: - **licenses** ([str]): associated licenses for that content - **tool** (dict): Tool used to compute the license """ ... @remote_api_endpoint('content/fossology_license/add') - def content_fossology_license_add(self, licenses, conflict_update=False): + def content_fossology_license_add( + self, licenses: List[Dict], conflict_update: bool = False) -> Dict: """Add licenses not present in storage. Args: licenses (iterable): dictionaries with keys: - **id**: sha1 - **licenses** ([bytes]): List of licenses associated to sha1 - **tool** (str): nomossa conflict_update: Flag to determine if we want to overwrite (true) or skip duplicates (false, the default) Returns: - list: content_license entries which failed due to unknown licenses + Dict summary of number of rows added """ ... @remote_api_endpoint('content/fossology_license/range') def content_fossology_license_get_range( self, start, end, indexer_configuration_id, limit=1000): """Retrieve licenses within range [start, end] bound by limit. Args: **start** (bytes): Starting identifier range (expected smaller than end) **end** (bytes): Ending identifier range (expected larger than start) **indexer_configuration_id** (int): The tool used to index data **limit** (int): Limit result (default to 1000) Raises: ValueError for limit to None Returns: a dict with keys: - **ids** [bytes]: iterable of content ids within the range. - **next** (Optional[bytes]): The next range of sha1 starts at this sha1 if any """ ... @remote_api_endpoint('content_metadata/missing') def content_metadata_missing(self, metadata): """List metadata missing from storage. Args: metadata (iterable): dictionaries with keys: - **id** (bytes): sha1 identifier - **indexer_configuration_id** (int): tool used to compute the results Yields: missing sha1s """ ... @remote_api_endpoint('content_metadata') def content_metadata_get(self, ids): """Retrieve metadata per id. Args: ids (iterable): sha1 checksums Yields: dictionaries with the following keys: id (bytes) metadata (str): associated metadata tool (dict): tool used to compute metadata """ ... @remote_api_endpoint('content_metadata/add') - def content_metadata_add(self, metadata, conflict_update=False): + def content_metadata_add( + self, metadata: List[Dict], conflict_update: bool = False) -> Dict: """Add metadata not present in storage. Args: metadata (iterable): dictionaries with keys: - **id**: sha1 - **metadata**: arbitrary dict conflict_update: Flag to determine if we want to overwrite (true) or skip duplicates (false, the default) + Returns: + Dict summary of number of rows added + """ ... @remote_api_endpoint('revision_intrinsic_metadata/missing') def revision_intrinsic_metadata_missing(self, metadata): """List metadata missing from storage. Args: metadata (iterable): dictionaries with keys: - **id** (bytes): sha1_git revision identifier - **indexer_configuration_id** (int): tool used to compute the results Yields: missing ids """ ... @remote_api_endpoint('revision_intrinsic_metadata') def revision_intrinsic_metadata_get(self, ids): """Retrieve revision metadata per id. Args: ids (iterable): sha1 checksums Yields: : dictionaries with the following keys: - **id** (bytes) - **metadata** (str): associated metadata - **tool** (dict): tool used to compute metadata - **mappings** (List[str]): list of mappings used to translate these metadata """ ... @remote_api_endpoint('revision_intrinsic_metadata/add') - def revision_intrinsic_metadata_add(self, metadata, conflict_update=False): + def revision_intrinsic_metadata_add( + self, metadata: List[Dict], conflict_update: bool = False) -> Dict: """Add metadata not present in storage. Args: metadata (iterable): dictionaries with keys: - **id**: sha1_git of revision - **metadata**: arbitrary dict - **indexer_configuration_id**: tool used to compute metadata - **mappings** (List[str]): list of mappings used to translate these metadata conflict_update: Flag to determine if we want to overwrite (true) or skip duplicates (false, the default) + Returns: + Dict summary of number of rows added + """ ... @remote_api_endpoint('revision_intrinsic_metadata/delete') - def revision_intrinsic_metadata_delete(self, entries): + def revision_intrinsic_metadata_delete(self, entries: List[Dict]) -> Dict: """Remove revision metadata from the storage. Args: entries (dict): dictionaries with the following keys: - **id** (bytes): revision identifier - **indexer_configuration_id** (int): tool used to compute metadata + + Returns: + Summary of number of rows deleted """ ... @remote_api_endpoint('origin_intrinsic_metadata') def origin_intrinsic_metadata_get(self, ids): """Retrieve origin metadata per id. Args: ids (iterable): origin identifiers Yields: list: dictionaries with the following keys: - **id** (str): origin url - **from_revision** (bytes): which revision this metadata was extracted from - **metadata** (str): associated metadata - **tool** (dict): tool used to compute metadata - **mappings** (List[str]): list of mappings used to translate these metadata """ ... @remote_api_endpoint('origin_intrinsic_metadata/add') - def origin_intrinsic_metadata_add(self, metadata, - conflict_update=False): + def origin_intrinsic_metadata_add( + self, metadata: List[Dict], conflict_update: bool = False) -> Dict: """Add origin metadata not present in storage. Args: metadata (iterable): dictionaries with keys: - **id**: origin urls - **from_revision**: sha1 id of the revision used to generate these metadata. - **metadata**: arbitrary dict - **indexer_configuration_id**: tool used to compute metadata - **mappings** (List[str]): list of mappings used to translate these metadata conflict_update: Flag to determine if we want to overwrite (true) or skip duplicates (false, the default) + Returns: + Dict summary of number of rows added + """ ... @remote_api_endpoint('origin_intrinsic_metadata/delete') def origin_intrinsic_metadata_delete( - self, entries): + self, entries: List[Dict]) -> Dict: """Remove origin metadata from the storage. Args: entries (dict): dictionaries with the following keys: - **id** (str): origin urls - **indexer_configuration_id** (int): tool used to compute metadata + + Returns: + Summary of number of rows deleted """ ... @remote_api_endpoint('origin_intrinsic_metadata/search/fulltext') def origin_intrinsic_metadata_search_fulltext( self, conjunction, limit=100): """Returns the list of origins whose metadata contain all the terms. Args: conjunction (List[str]): List of terms to be searched for. limit (int): The maximum number of results to return Yields: list: dictionaries with the following keys: - **id** (str): origin urls - **from_revision**: sha1 id of the revision used to generate these metadata. - **metadata** (str): associated metadata - **tool** (dict): tool used to compute metadata - **mappings** (List[str]): list of mappings used to translate these metadata """ ... @remote_api_endpoint('origin_intrinsic_metadata/search/by_producer') def origin_intrinsic_metadata_search_by_producer( self, page_token='', limit=100, ids_only=False, mappings=None, tool_ids=None): """Returns the list of origins whose metadata contain all the terms. Args: page_token (str): Opaque token used for pagination. limit (int): The maximum number of results to return ids_only (bool): Determines whether only origin urls are returned or the content as well mappings (List[str]): Returns origins whose intrinsic metadata were generated using at least one of these mappings. Returns: dict: dict with the following keys: - **next_page_token** (str, optional): opaque token to be used as `page_token` for retrieving the next page. If absent, there is no more pages to gather. - **origins** (list): list of origin url (str) if `ids_only=True` else dictionaries with the following keys: - **id** (str): origin urls - **from_revision**: sha1 id of the revision used to generate these metadata. - **metadata** (str): associated metadata - **tool** (dict): tool used to compute metadata - **mappings** (List[str]): list of mappings used to translate these metadata """ ... @remote_api_endpoint('origin_intrinsic_metadata/stats') def origin_intrinsic_metadata_stats( self): """Returns counts of indexed metadata per origins, broken down into metadata types. Returns: dict: dictionary with keys: - total (int): total number of origins that were indexed (possibly yielding an empty metadata dictionary) - non_empty (int): total number of origins that we extracted a non-empty metadata dictionary from - per_mapping (dict): a dictionary with mapping names as keys and number of origins whose indexing used this mapping. Note that indexing a given origin may use 0, 1, or many mappings. """ ... @remote_api_endpoint('indexer_configuration/add') def indexer_configuration_add(self, tools): """Add new tools to the storage. Args: tools ([dict]): List of dictionary representing tool to insert in the db. Dictionary with the following keys: - **tool_name** (str): tool's name - **tool_version** (str): tool's version - **tool_configuration** (dict): tool's configuration (free form dict) Returns: List of dict inserted in the db (holding the id key as well). The order of the list is not guaranteed to match the order of the initial list. """ ... @remote_api_endpoint('indexer_configuration/data') def indexer_configuration_get(self, tool): """Retrieve tool information. Args: tool (dict): Dictionary representing a tool with the following keys: - **tool_name** (str): tool's name - **tool_version** (str): tool's version - **tool_configuration** (dict): tool's configuration (free form dict) Returns: The same dictionary with an `id` key, None otherwise. """ ... diff --git a/swh/indexer/tests/storage/test_storage.py b/swh/indexer/tests/storage/test_storage.py index 8f4c2fb..8c3d994 100644 --- a/swh/indexer/tests/storage/test_storage.py +++ b/swh/indexer/tests/storage/test_storage.py @@ -1,1889 +1,1933 @@ -# Copyright (C) 2015-2019 The Software Heritage developers +# Copyright (C) 2015-2020 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import inspect import threading +from typing import Dict + import pytest from swh.model.hashutil import hash_to_bytes from swh.indexer.storage.exc import ( IndexerStorageArgumentException, DuplicateId, ) from swh.indexer.storage.interface import IndexerStorageInterface def prepare_mimetypes_from(fossology_licenses): """Fossology license needs some consistent data in db to run. """ mimetypes = [] for c in fossology_licenses: mimetypes.append({ 'id': c['id'], 'mimetype': 'text/plain', 'encoding': 'utf-8', 'indexer_configuration_id': c['indexer_configuration_id'], }) return mimetypes -def endpoint(storage, endpoint_type, endpoint_name): - return getattr(storage, endpoint_type + '_' + endpoint_name) +def endpoint_name(etype: str, ename: str) -> str: + """Compute the storage's endpoint's name + + >>> endpoint_name('content_mimetype', 'add') + 'content_mimetype_add' + >>> endpoint_name('content_fosso_license', 'delete') + 'content_fosso_license_delete' + + """ + return f'{etype}_{ename}' + + +def endpoint(storage, etype: str, ename: str): + return getattr(storage, endpoint_name(etype, ename)) + + +def expected_summary( + count: int, etype: str, ename: str = 'add') -> Dict[str, int]: + """Compute the expected summary + + The key is determine according to etype and ename + + >>> expected_summary(10, 'content_mimetype', 'add') + {'content_mimetype:add': 10} + >>> expected_summary(9, 'origin_intrinsic_metadata', 'delete') + {'origin_intrinsic_metadata:del': 9} + + """ + pattern = ename[0:3] + key = endpoint_name(etype, ename).replace(f'_{ename}', f':{pattern}') + return { + key: count + } class StorageETypeTester: """Base class for testing a series of common behaviour between a bunch of endpoint types supported by an IndexerStorage. This is supposed to be inherited with the following class attributes: - endpoint_type - tool_name - example_data See below for example usage. """ def test_missing(self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data etype = self.endpoint_type tool_id = data.tools[self.tool_name]['id'] # given 2 (hopefully) unknown objects query = [ { 'id': data.sha1_1, 'indexer_configuration_id': tool_id, }, { 'id': data.sha1_2, 'indexer_configuration_id': tool_id, }] # we expect these are both returned by the xxx_missing endpoint actual_missing = endpoint(storage, etype, 'missing')(query) assert list(actual_missing) == [ data.sha1_1, data.sha1_2, ] # now, when we add one of them - endpoint(storage, etype, 'add')([{ + summary = endpoint(storage, etype, 'add')([{ 'id': data.sha1_2, **self.example_data[0], 'indexer_configuration_id': tool_id, }]) + assert summary == expected_summary(1, etype) + # we expect only the other one returned actual_missing = endpoint(storage, etype, 'missing')(query) assert list(actual_missing) == [data.sha1_1] def test_add__drop_duplicate(self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data etype = self.endpoint_type tool_id = data.tools[self.tool_name]['id'] # add the first object data_v1 = { 'id': data.sha1_2, **self.example_data[0], 'indexer_configuration_id': tool_id, } - endpoint(storage, etype, 'add')([data_v1]) + summary = endpoint(storage, etype, 'add')([data_v1]) + assert summary == expected_summary(1, etype) # should be able to retrieve it actual_data = list(endpoint(storage, etype, 'get')([data.sha1_2])) expected_data_v1 = [{ 'id': data.sha1_2, **self.example_data[0], 'tool': data.tools[self.tool_name], }] assert actual_data == expected_data_v1 # now if we add a modified version of the same object (same id) data_v2 = data_v1.copy() data_v2.update(self.example_data[1]) - endpoint(storage, etype, 'add')([data_v2]) + summary2 = endpoint(storage, etype, 'add')([data_v2]) + assert summary2 == expected_summary(0, etype) # not added # we expect to retrieve the original data, not the modified one actual_data = list(endpoint(storage, etype, 'get')([data.sha1_2])) assert actual_data == expected_data_v1 def test_add__update_in_place_duplicate( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data etype = self.endpoint_type tool = data.tools[self.tool_name] data_v1 = { 'id': data.sha1_2, **self.example_data[0], 'indexer_configuration_id': tool['id'], } # given - endpoint(storage, etype, 'add')([data_v1]) + summary = endpoint(storage, etype, 'add')([data_v1]) + assert summary == expected_summary(1, etype) # not added # when actual_data = list(endpoint(storage, etype, 'get')([data.sha1_2])) expected_data_v1 = [{ 'id': data.sha1_2, **self.example_data[0], 'tool': tool, }] # then assert actual_data == expected_data_v1 # given data_v2 = data_v1.copy() data_v2.update(self.example_data[1]) endpoint(storage, etype, 'add')([data_v2], conflict_update=True) + assert summary == expected_summary(1, etype) # modified so counted actual_data = list(endpoint(storage, etype, 'get')([data.sha1_2])) expected_data_v2 = [{ 'id': data.sha1_2, **self.example_data[1], 'tool': tool, }] # data did change as the v2 was used to overwrite v1 assert actual_data == expected_data_v2 def test_add__update_in_place_deadlock( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data etype = self.endpoint_type tool = data.tools[self.tool_name] hashes = [ hash_to_bytes( '34973274ccef6ab4dfaaf86599792fa9c3fe4{:03d}'.format(i)) for i in range(1000)] data_v1 = [ { 'id': hash_, **self.example_data[0], 'indexer_configuration_id': tool['id'], } for hash_ in hashes ] data_v2 = [ { 'id': hash_, **self.example_data[1], 'indexer_configuration_id': tool['id'], } for hash_ in hashes ] # Remove one item from each, so that both queries have to succeed for # all items to be in the DB. data_v2a = data_v2[1:] data_v2b = list(reversed(data_v2[0:-1])) # given endpoint(storage, etype, 'add')(data_v1) # when actual_data = list(endpoint(storage, etype, 'get')(hashes)) expected_data_v1 = [ { 'id': hash_, **self.example_data[0], 'tool': tool, } for hash_ in hashes ] # then assert actual_data == expected_data_v1 # given def f1(): endpoint(storage, etype, 'add')(data_v2a, conflict_update=True) def f2(): endpoint(storage, etype, 'add')(data_v2b, conflict_update=True) t1 = threading.Thread(target=f1) t2 = threading.Thread(target=f2) t2.start() t1.start() t1.join() t2.join() actual_data = sorted(endpoint(storage, etype, 'get')(hashes), key=lambda x: x['id']) expected_data_v2 = [ { 'id': hash_, **self.example_data[1], 'tool': tool, } for hash_ in hashes ] assert actual_data == expected_data_v2 def test_add__duplicate_twice(self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data etype = self.endpoint_type tool = data.tools[self.tool_name] data_rev1 = { 'id': data.revision_id_2, **self.example_data[0], 'indexer_configuration_id': tool['id'] } data_rev2 = { 'id': data.revision_id_2, **self.example_data[1], 'indexer_configuration_id': tool['id'] } # when - endpoint(storage, etype, 'add')([data_rev1]) + summary = endpoint(storage, etype, 'add')([data_rev1]) + assert summary == expected_summary(1, etype) with pytest.raises(DuplicateId): endpoint(storage, etype, 'add')( [data_rev2, data_rev2], conflict_update=True) # then actual_data = list(endpoint(storage, etype, 'get')( [data.revision_id_2, data.revision_id_1])) expected_data = [{ 'id': data.revision_id_2, **self.example_data[0], 'tool': tool, }] assert actual_data == expected_data def test_get(self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data etype = self.endpoint_type tool = data.tools[self.tool_name] query = [data.sha1_2, data.sha1_1] data1 = { 'id': data.sha1_2, **self.example_data[0], 'indexer_configuration_id': tool['id'], } # when - endpoint(storage, etype, 'add')([data1]) + summary = endpoint(storage, etype, 'add')([data1]) + assert summary == expected_summary(1, etype) # then actual_data = list(endpoint(storage, etype, 'get')(query)) # then expected_data = [{ 'id': data.sha1_2, **self.example_data[0], 'tool': tool, }] assert actual_data == expected_data class TestIndexerStorageContentMimetypes(StorageETypeTester): """Test Indexer Storage content_mimetype related methods """ endpoint_type = 'content_mimetype' tool_name = 'file' example_data = [ { 'mimetype': 'text/plain', 'encoding': 'utf-8', }, { 'mimetype': 'text/html', 'encoding': 'us-ascii', }, ] def test_generate_content_mimetype_get_range_limit_none( self, swh_indexer_storage): """mimetype_get_range call with wrong limit input should fail""" storage = swh_indexer_storage with pytest.raises(IndexerStorageArgumentException) as e: storage.content_mimetype_get_range( start=None, end=None, indexer_configuration_id=None, limit=None) assert e.value.args == ('limit should not be None',) def test_generate_content_mimetype_get_range_no_limit( self, swh_indexer_storage_with_data): """mimetype_get_range returns mimetypes within range provided""" storage, data = swh_indexer_storage_with_data mimetypes = data.mimetypes # All ids from the db content_ids = sorted([c['id'] for c in mimetypes]) start = content_ids[0] end = content_ids[-1] # retrieve mimetypes tool_id = mimetypes[0]['indexer_configuration_id'] actual_result = storage.content_mimetype_get_range( start, end, indexer_configuration_id=tool_id) actual_ids = actual_result['ids'] actual_next = actual_result['next'] assert len(mimetypes) == len(actual_ids) assert actual_next is None assert content_ids == actual_ids def test_generate_content_mimetype_get_range_limit( self, swh_indexer_storage_with_data): """mimetype_get_range paginates results if limit exceeded""" storage, data = swh_indexer_storage_with_data indexer_configuration_id = data.tools['file']['id'] # input the list of sha1s we want from storage content_ids = sorted( [c['id'] for c in data.mimetypes]) mimetypes = list(storage.content_mimetype_get(content_ids)) assert len(mimetypes) == len(data.mimetypes) start = content_ids[0] end = content_ids[-1] # retrieve mimetypes limited to 10 results actual_result = storage.content_mimetype_get_range( start, end, indexer_configuration_id=indexer_configuration_id, limit=10) assert actual_result assert set(actual_result.keys()) == {'ids', 'next'} actual_ids = actual_result['ids'] actual_next = actual_result['next'] assert len(actual_ids) == 10 assert actual_next is not None assert actual_next == content_ids[10] expected_mimetypes = content_ids[:10] assert expected_mimetypes == actual_ids # retrieve next part actual_result = storage.content_mimetype_get_range( start=end, end=end, indexer_configuration_id=indexer_configuration_id) assert set(actual_result.keys()) == {'ids', 'next'} actual_ids = actual_result['ids'] actual_next = actual_result['next'] assert actual_next is None expected_mimetypes = [content_ids[-1]] assert expected_mimetypes == actual_ids class TestIndexerStorageContentLanguage(StorageETypeTester): """Test Indexer Storage content_language related methods """ endpoint_type = 'content_language' tool_name = 'pygments' example_data = [ { 'lang': 'haskell', }, { 'lang': 'common-lisp', }, ] class TestIndexerStorageContentCTags(StorageETypeTester): """Test Indexer Storage content_ctags related methods """ endpoint_type = 'content_ctags' tool_name = 'universal-ctags' example_data = [ { 'ctags': [{ 'name': 'done', 'kind': 'variable', 'line': 119, 'lang': 'OCaml', }] }, { 'ctags': [ { 'name': 'done', 'kind': 'variable', 'line': 100, 'lang': 'Python', }, { 'name': 'main', 'kind': 'function', 'line': 119, 'lang': 'Python', }] }, ] # the following tests are disabled because CTAGS behaves differently @pytest.mark.skip def test_add__drop_duplicate(self): pass @pytest.mark.skip def test_add__update_in_place_duplicate(self): pass @pytest.mark.skip def test_add__update_in_place_deadlock(self): pass @pytest.mark.skip def test_add__duplicate_twice(self): pass @pytest.mark.skip def test_get(self): pass def test_content_ctags_search(self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data # 1. given tool = data.tools['universal-ctags'] tool_id = tool['id'] ctag1 = { 'id': data.sha1_1, 'indexer_configuration_id': tool_id, 'ctags': [ { 'name': 'hello', 'kind': 'function', 'line': 133, 'lang': 'Python', }, { 'name': 'counter', 'kind': 'variable', 'line': 119, 'lang': 'Python', }, { 'name': 'hello', 'kind': 'variable', 'line': 210, 'lang': 'Python', }, ] } ctag2 = { 'id': data.sha1_2, 'indexer_configuration_id': tool_id, 'ctags': [ { 'name': 'hello', 'kind': 'variable', 'line': 100, 'lang': 'C', }, { 'name': 'result', 'kind': 'variable', 'line': 120, 'lang': 'C', }, ] } storage.content_ctags_add([ctag1, ctag2]) # 1. when actual_ctags = list(storage.content_ctags_search('hello', limit=1)) # 1. then assert actual_ctags == [ { 'id': ctag1['id'], 'tool': tool, 'name': 'hello', 'kind': 'function', 'line': 133, 'lang': 'Python', } ] # 2. when actual_ctags = list(storage.content_ctags_search( 'hello', limit=1, last_sha1=ctag1['id'])) # 2. then assert actual_ctags == [ { 'id': ctag2['id'], 'tool': tool, 'name': 'hello', 'kind': 'variable', 'line': 100, 'lang': 'C', } ] # 3. when actual_ctags = list(storage.content_ctags_search('hello')) # 3. then assert actual_ctags == [ { 'id': ctag1['id'], 'tool': tool, 'name': 'hello', 'kind': 'function', 'line': 133, 'lang': 'Python', }, { 'id': ctag1['id'], 'tool': tool, 'name': 'hello', 'kind': 'variable', 'line': 210, 'lang': 'Python', }, { 'id': ctag2['id'], 'tool': tool, 'name': 'hello', 'kind': 'variable', 'line': 100, 'lang': 'C', }, ] # 4. when actual_ctags = list(storage.content_ctags_search('counter')) # then assert actual_ctags == [{ 'id': ctag1['id'], 'tool': tool, 'name': 'counter', 'kind': 'variable', 'line': 119, 'lang': 'Python', }] # 5. when actual_ctags = list(storage.content_ctags_search('result', limit=1)) # then assert actual_ctags == [{ 'id': ctag2['id'], 'tool': tool, 'name': 'result', 'kind': 'variable', 'line': 120, 'lang': 'C', }] def test_content_ctags_search_no_result(self, swh_indexer_storage): storage = swh_indexer_storage actual_ctags = list(storage.content_ctags_search('counter')) assert not actual_ctags def test_content_ctags_add__add_new_ctags_added( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data # given tool = data.tools['universal-ctags'] tool_id = tool['id'] ctag_v1 = { 'id': data.sha1_2, 'indexer_configuration_id': tool_id, 'ctags': [{ 'name': 'done', 'kind': 'variable', 'line': 100, 'lang': 'Scheme', }] } # given storage.content_ctags_add([ctag_v1]) storage.content_ctags_add([ctag_v1]) # conflict does nothing # when actual_ctags = list(storage.content_ctags_get([data.sha1_2])) # then expected_ctags = [{ 'id': data.sha1_2, 'name': 'done', 'kind': 'variable', 'line': 100, 'lang': 'Scheme', 'tool': tool, }] assert actual_ctags == expected_ctags # given ctag_v2 = ctag_v1.copy() ctag_v2.update({ 'ctags': [ { 'name': 'defn', 'kind': 'function', 'line': 120, 'lang': 'Scheme', } ] }) storage.content_ctags_add([ctag_v2]) expected_ctags = [ { 'id': data.sha1_2, 'name': 'done', 'kind': 'variable', 'line': 100, 'lang': 'Scheme', 'tool': tool, }, { 'id': data.sha1_2, 'name': 'defn', 'kind': 'function', 'line': 120, 'lang': 'Scheme', 'tool': tool, } ] actual_ctags = list(storage.content_ctags_get( [data.sha1_2])) assert actual_ctags == expected_ctags def test_content_ctags_add__update_in_place( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data # given tool = data.tools['universal-ctags'] tool_id = tool['id'] ctag_v1 = { 'id': data.sha1_2, 'indexer_configuration_id': tool_id, 'ctags': [{ 'name': 'done', 'kind': 'variable', 'line': 100, 'lang': 'Scheme', }] } # given storage.content_ctags_add([ctag_v1]) # when actual_ctags = list(storage.content_ctags_get( [data.sha1_2])) # then expected_ctags = [ { 'id': data.sha1_2, 'name': 'done', 'kind': 'variable', 'line': 100, 'lang': 'Scheme', 'tool': tool } ] assert actual_ctags == expected_ctags # given ctag_v2 = ctag_v1.copy() ctag_v2.update({ 'ctags': [ { 'name': 'done', 'kind': 'variable', 'line': 100, 'lang': 'Scheme', }, { 'name': 'defn', 'kind': 'function', 'line': 120, 'lang': 'Scheme', } ] }) storage.content_ctags_add([ctag_v2], conflict_update=True) actual_ctags = list(storage.content_ctags_get( [data.sha1_2])) # ctag did change as the v2 was used to overwrite v1 expected_ctags = [ { 'id': data.sha1_2, 'name': 'done', 'kind': 'variable', 'line': 100, 'lang': 'Scheme', 'tool': tool, }, { 'id': data.sha1_2, 'name': 'defn', 'kind': 'function', 'line': 120, 'lang': 'Scheme', 'tool': tool, } ] assert actual_ctags == expected_ctags class TestIndexerStorageContentMetadata(StorageETypeTester): """Test Indexer Storage content_metadata related methods """ tool_name = 'swh-metadata-detector' endpoint_type = 'content_metadata' example_data = [ { 'metadata': { 'other': {}, 'codeRepository': { 'type': 'git', 'url': 'https://github.com/moranegg/metadata_test' }, 'description': 'Simple package.json test for indexer', 'name': 'test_metadata', 'version': '0.0.1' }, }, { 'metadata': { 'other': {}, 'name': 'test_metadata', 'version': '0.0.1' }, }, ] class TestIndexerStorageRevisionIntrinsicMetadata(StorageETypeTester): """Test Indexer Storage revision_intrinsic_metadata related methods """ tool_name = 'swh-metadata-detector' endpoint_type = 'revision_intrinsic_metadata' example_data = [ { 'metadata': { 'other': {}, 'codeRepository': { 'type': 'git', 'url': 'https://github.com/moranegg/metadata_test' }, 'description': 'Simple package.json test for indexer', 'name': 'test_metadata', 'version': '0.0.1' }, 'mappings': ['mapping1'], }, { 'metadata': { 'other': {}, 'name': 'test_metadata', 'version': '0.0.1' }, 'mappings': ['mapping2'], }, ] def test_revision_intrinsic_metadata_delete( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data etype = self.endpoint_type tool = data.tools[self.tool_name] query = [data.sha1_2, data.sha1_1] data1 = { 'id': data.sha1_2, **self.example_data[0], 'indexer_configuration_id': tool['id'], } # when - endpoint(storage, etype, 'add')([data1]) - endpoint(storage, etype, 'delete')([ + summary = endpoint(storage, etype, 'add')([data1]) + assert summary == expected_summary(1, etype) + + summary2 = endpoint(storage, etype, 'delete')([ { 'id': data.sha1_2, 'indexer_configuration_id': tool['id'], } ]) + assert summary2 == expected_summary(1, etype, 'del') # then actual_data = list(endpoint(storage, etype, 'get')(query)) # then assert not actual_data def test_revision_intrinsic_metadata_delete_nonexisting( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data etype = self.endpoint_type tool = data.tools[self.tool_name] endpoint(storage, etype, 'delete')([ { 'id': data.sha1_2, 'indexer_configuration_id': tool['id'], } ]) class TestIndexerStorageContentFossologyLicence: def test_content_fossology_license_add__new_license_added( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data # given tool = data.tools['nomos'] tool_id = tool['id'] license_v1 = { 'id': data.sha1_1, 'licenses': ['Apache-2.0'], 'indexer_configuration_id': tool_id, } # given storage.content_fossology_license_add([license_v1]) # conflict does nothing storage.content_fossology_license_add([license_v1]) # when actual_licenses = list(storage.content_fossology_license_get( [data.sha1_1])) # then expected_license = { data.sha1_1: [{ 'licenses': ['Apache-2.0'], 'tool': tool, }] } assert actual_licenses == [expected_license] # given license_v2 = license_v1.copy() license_v2.update({ 'licenses': ['BSD-2-Clause'], }) storage.content_fossology_license_add([license_v2]) actual_licenses = list(storage.content_fossology_license_get( [data.sha1_1])) expected_license = { data.sha1_1: [{ 'licenses': ['Apache-2.0', 'BSD-2-Clause'], 'tool': tool }] } # license did not change as the v2 was dropped. assert actual_licenses == [expected_license] def test_generate_content_fossology_license_get_range_limit_none( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data """license_get_range call with wrong limit input should fail""" with pytest.raises(IndexerStorageArgumentException) as e: storage.content_fossology_license_get_range( start=None, end=None, indexer_configuration_id=None, limit=None) assert e.value.args == ('limit should not be None',) def test_generate_content_fossology_license_get_range_no_limit( self, swh_indexer_storage_with_data): """license_get_range returns licenses within range provided""" storage, data = swh_indexer_storage_with_data # craft some consistent mimetypes fossology_licenses = data.fossology_licenses mimetypes = prepare_mimetypes_from(fossology_licenses) storage.content_mimetype_add(mimetypes, conflict_update=True) # add fossology_licenses to storage storage.content_fossology_license_add(fossology_licenses) # All ids from the db content_ids = sorted([c['id'] for c in fossology_licenses]) start = content_ids[0] end = content_ids[-1] # retrieve fossology_licenses tool_id = fossology_licenses[0]['indexer_configuration_id'] actual_result = storage.content_fossology_license_get_range( start, end, indexer_configuration_id=tool_id) actual_ids = actual_result['ids'] actual_next = actual_result['next'] assert len(fossology_licenses) == len(actual_ids) assert actual_next is None assert content_ids == actual_ids def test_generate_content_fossology_license_get_range_no_limit_with_filter( self, swh_indexer_storage_with_data): """This filters non textual, then returns results within range""" storage, data = swh_indexer_storage_with_data fossology_licenses = data.fossology_licenses mimetypes = data.mimetypes # craft some consistent mimetypes _mimetypes = prepare_mimetypes_from(fossology_licenses) # add binary mimetypes which will get filtered out in results for m in mimetypes: _mimetypes.append({ 'mimetype': 'binary', **m, }) storage.content_mimetype_add(_mimetypes, conflict_update=True) # add fossology_licenses to storage storage.content_fossology_license_add(fossology_licenses) # All ids from the db content_ids = sorted([c['id'] for c in fossology_licenses]) start = content_ids[0] end = content_ids[-1] # retrieve fossology_licenses tool_id = fossology_licenses[0]['indexer_configuration_id'] actual_result = storage.content_fossology_license_get_range( start, end, indexer_configuration_id=tool_id) actual_ids = actual_result['ids'] actual_next = actual_result['next'] assert len(fossology_licenses) == len(actual_ids) assert actual_next is None assert content_ids == actual_ids def test_generate_fossology_license_get_range_limit( self, swh_indexer_storage_with_data): """fossology_license_get_range paginates results if limit exceeded""" storage, data = swh_indexer_storage_with_data fossology_licenses = data.fossology_licenses # craft some consistent mimetypes mimetypes = prepare_mimetypes_from(fossology_licenses) # add fossology_licenses to storage storage.content_mimetype_add(mimetypes, conflict_update=True) storage.content_fossology_license_add(fossology_licenses) # input the list of sha1s we want from storage content_ids = sorted([c['id'] for c in fossology_licenses]) start = content_ids[0] end = content_ids[-1] # retrieve fossology_licenses limited to 3 results limited_results = len(fossology_licenses) - 1 tool_id = fossology_licenses[0]['indexer_configuration_id'] actual_result = storage.content_fossology_license_get_range( start, end, indexer_configuration_id=tool_id, limit=limited_results) actual_ids = actual_result['ids'] actual_next = actual_result['next'] assert limited_results == len(actual_ids) assert actual_next is not None assert actual_next == content_ids[-1] expected_fossology_licenses = content_ids[:-1] assert expected_fossology_licenses == actual_ids # retrieve next part actual_results2 = storage.content_fossology_license_get_range( start=end, end=end, indexer_configuration_id=tool_id) actual_ids2 = actual_results2['ids'] actual_next2 = actual_results2['next'] assert actual_next2 is None expected_fossology_licenses2 = [content_ids[-1]] assert expected_fossology_licenses2 == actual_ids2 class TestIndexerStorageOriginIntrinsicMetadata: def test_origin_intrinsic_metadata_get( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data # given tool_id = data.tools['swh-metadata-detector']['id'] metadata = { 'version': None, 'name': None, } metadata_rev = { 'id': data.revision_id_2, 'metadata': metadata, 'mappings': ['mapping1'], 'indexer_configuration_id': tool_id, } metadata_origin = { 'id': data.origin_url_1, 'metadata': metadata, 'indexer_configuration_id': tool_id, 'mappings': ['mapping1'], 'from_revision': data.revision_id_2, } # when storage.revision_intrinsic_metadata_add([metadata_rev]) storage.origin_intrinsic_metadata_add([metadata_origin]) # then actual_metadata = list(storage.origin_intrinsic_metadata_get( [data.origin_url_1, 'no://where'])) expected_metadata = [{ 'id': data.origin_url_1, 'metadata': metadata, 'tool': data.tools['swh-metadata-detector'], 'from_revision': data.revision_id_2, 'mappings': ['mapping1'], }] assert actual_metadata == expected_metadata def test_origin_intrinsic_metadata_delete( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data # given tool_id = data.tools['swh-metadata-detector']['id'] metadata = { 'version': None, 'name': None, } metadata_rev = { 'id': data.revision_id_2, 'metadata': metadata, 'mappings': ['mapping1'], 'indexer_configuration_id': tool_id, } metadata_origin = { 'id': data.origin_url_1, 'metadata': metadata, 'indexer_configuration_id': tool_id, 'mappings': ['mapping1'], 'from_revision': data.revision_id_2, } metadata_origin2 = metadata_origin.copy() metadata_origin2['id'] = data.origin_url_2 # when storage.revision_intrinsic_metadata_add([metadata_rev]) storage.origin_intrinsic_metadata_add([ metadata_origin, metadata_origin2]) storage.origin_intrinsic_metadata_delete([ { 'id': data.origin_url_1, 'indexer_configuration_id': tool_id } ]) # then actual_metadata = list(storage.origin_intrinsic_metadata_get( [data.origin_url_1, data.origin_url_2, 'no://where'])) for item in actual_metadata: item['indexer_configuration_id'] = item.pop('tool')['id'] assert actual_metadata == [metadata_origin2] def test_origin_intrinsic_metadata_delete_nonexisting( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data tool_id = data.tools['swh-metadata-detector']['id'] storage.origin_intrinsic_metadata_delete([ { 'id': data.origin_url_1, 'indexer_configuration_id': tool_id } ]) def test_origin_intrinsic_metadata_add_drop_duplicate( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data # given tool_id = data.tools['swh-metadata-detector']['id'] metadata_v1 = { 'version': None, 'name': None, } metadata_rev_v1 = { 'id': data.revision_id_1, 'metadata': metadata_v1.copy(), 'mappings': [], 'indexer_configuration_id': tool_id, } metadata_origin_v1 = { 'id': data.origin_url_1, 'metadata': metadata_v1.copy(), 'indexer_configuration_id': tool_id, 'mappings': [], 'from_revision': data.revision_id_1, } # given storage.revision_intrinsic_metadata_add([metadata_rev_v1]) storage.origin_intrinsic_metadata_add([metadata_origin_v1]) # when actual_metadata = list(storage.origin_intrinsic_metadata_get( [data.origin_url_1, 'no://where'])) expected_metadata_v1 = [{ 'id': data.origin_url_1, 'metadata': metadata_v1, 'tool': data.tools['swh-metadata-detector'], 'from_revision': data.revision_id_1, 'mappings': [], }] assert actual_metadata == expected_metadata_v1 # given metadata_v2 = metadata_v1.copy() metadata_v2.update({ 'name': 'test_metadata', 'author': 'MG', }) metadata_rev_v2 = metadata_rev_v1.copy() metadata_origin_v2 = metadata_origin_v1.copy() metadata_rev_v2['metadata'] = metadata_v2 metadata_origin_v2['metadata'] = metadata_v2 storage.revision_intrinsic_metadata_add([metadata_rev_v2]) storage.origin_intrinsic_metadata_add([metadata_origin_v2]) # then actual_metadata = list(storage.origin_intrinsic_metadata_get( [data.origin_url_1])) # metadata did not change as the v2 was dropped. assert actual_metadata == expected_metadata_v1 def test_origin_intrinsic_metadata_add_update_in_place_duplicate( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data # given tool_id = data.tools['swh-metadata-detector']['id'] metadata_v1 = { 'version': None, 'name': None, } metadata_rev_v1 = { 'id': data.revision_id_2, 'metadata': metadata_v1, 'mappings': [], 'indexer_configuration_id': tool_id, } metadata_origin_v1 = { 'id': data.origin_url_1, 'metadata': metadata_v1.copy(), 'indexer_configuration_id': tool_id, 'mappings': [], 'from_revision': data.revision_id_2, } # given storage.revision_intrinsic_metadata_add([metadata_rev_v1]) storage.origin_intrinsic_metadata_add([metadata_origin_v1]) # when actual_metadata = list(storage.origin_intrinsic_metadata_get( [data.origin_url_1])) # then expected_metadata_v1 = [{ 'id': data.origin_url_1, 'metadata': metadata_v1, 'tool': data.tools['swh-metadata-detector'], 'from_revision': data.revision_id_2, 'mappings': [], }] assert actual_metadata == expected_metadata_v1 # given metadata_v2 = metadata_v1.copy() metadata_v2.update({ 'name': 'test_update_duplicated_metadata', 'author': 'MG', }) metadata_rev_v2 = metadata_rev_v1.copy() metadata_origin_v2 = metadata_origin_v1.copy() metadata_rev_v2['metadata'] = metadata_v2 metadata_origin_v2 = { 'id': data.origin_url_1, 'metadata': metadata_v2.copy(), 'indexer_configuration_id': tool_id, 'mappings': ['npm'], 'from_revision': data.revision_id_1, } storage.revision_intrinsic_metadata_add( [metadata_rev_v2], conflict_update=True) storage.origin_intrinsic_metadata_add( [metadata_origin_v2], conflict_update=True) actual_metadata = list(storage.origin_intrinsic_metadata_get( [data.origin_url_1])) expected_metadata_v2 = [{ 'id': data.origin_url_1, 'metadata': metadata_v2, 'tool': data.tools['swh-metadata-detector'], 'from_revision': data.revision_id_1, 'mappings': ['npm'], }] # metadata did change as the v2 was used to overwrite v1 assert actual_metadata == expected_metadata_v2 def test_origin_intrinsic_metadata_add__update_in_place_deadlock( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data # given tool_id = data.tools['swh-metadata-detector']['id'] ids = list(range(10)) example_data1 = { 'metadata': { 'version': None, 'name': None, }, 'mappings': [], } example_data2 = { 'metadata': { 'version': 'v1.1.1', 'name': 'foo', }, 'mappings': [], } metadata_rev_v1 = { 'id': data.revision_id_2, 'metadata': { 'version': None, 'name': None, }, 'mappings': [], 'indexer_configuration_id': tool_id, } data_v1 = [ { 'id': 'file:///tmp/origin%d' % id_, 'from_revision': data.revision_id_2, **example_data1, 'indexer_configuration_id': tool_id, } for id_ in ids ] data_v2 = [ { 'id': 'file:///tmp/origin%d' % id_, 'from_revision': data.revision_id_2, **example_data2, 'indexer_configuration_id': tool_id, } for id_ in ids ] # Remove one item from each, so that both queries have to succeed for # all items to be in the DB. data_v2a = data_v2[1:] data_v2b = list(reversed(data_v2[0:-1])) # given storage.revision_intrinsic_metadata_add([metadata_rev_v1]) storage.origin_intrinsic_metadata_add(data_v1) # when origins = ['file:///tmp/origin%d' % i for i in ids] actual_data = list(storage.origin_intrinsic_metadata_get(origins)) expected_data_v1 = [ { 'id': 'file:///tmp/origin%d' % id_, 'from_revision': data.revision_id_2, **example_data1, 'tool': data.tools['swh-metadata-detector'], } for id_ in ids ] # then assert actual_data == expected_data_v1 # given def f1(): storage.origin_intrinsic_metadata_add( data_v2a, conflict_update=True) def f2(): storage.origin_intrinsic_metadata_add( data_v2b, conflict_update=True) t1 = threading.Thread(target=f1) t2 = threading.Thread(target=f2) t2.start() t1.start() t1.join() t2.join() actual_data = list(storage.origin_intrinsic_metadata_get(origins)) expected_data_v2 = [ { 'id': 'file:///tmp/origin%d' % id_, 'from_revision': data.revision_id_2, **example_data2, 'tool': data.tools['swh-metadata-detector'], } for id_ in ids ] assert len(actual_data) == len(expected_data_v2) assert sorted(actual_data, key=lambda x: x['id']) == expected_data_v2 def test_origin_intrinsic_metadata_add__duplicate_twice( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data # given tool_id = data.tools['swh-metadata-detector']['id'] metadata = { 'developmentStatus': None, 'name': None, } metadata_rev = { 'id': data.revision_id_2, 'metadata': metadata, 'mappings': ['mapping1'], 'indexer_configuration_id': tool_id, } metadata_origin = { 'id': data.origin_url_1, 'metadata': metadata, 'indexer_configuration_id': tool_id, 'mappings': ['mapping1'], 'from_revision': data.revision_id_2, } # when storage.revision_intrinsic_metadata_add([metadata_rev]) with pytest.raises(DuplicateId): storage.origin_intrinsic_metadata_add([ metadata_origin, metadata_origin]) def test_origin_intrinsic_metadata_search_fulltext( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data # given tool_id = data.tools['swh-metadata-detector']['id'] metadata1 = { 'author': 'John Doe', } metadata1_rev = { 'id': data.revision_id_1, 'metadata': metadata1, 'mappings': [], 'indexer_configuration_id': tool_id, } metadata1_origin = { 'id': data.origin_url_1, 'metadata': metadata1, 'mappings': [], 'indexer_configuration_id': tool_id, 'from_revision': data.revision_id_1, } metadata2 = { 'author': 'Jane Doe', } metadata2_rev = { 'id': data.revision_id_2, 'metadata': metadata2, 'mappings': [], 'indexer_configuration_id': tool_id, } metadata2_origin = { 'id': data.origin_url_2, 'metadata': metadata2, 'mappings': [], 'indexer_configuration_id': tool_id, 'from_revision': data.revision_id_2, } # when storage.revision_intrinsic_metadata_add([metadata1_rev]) storage.origin_intrinsic_metadata_add([metadata1_origin]) storage.revision_intrinsic_metadata_add([metadata2_rev]) storage.origin_intrinsic_metadata_add([metadata2_origin]) # then search = storage.origin_intrinsic_metadata_search_fulltext assert set([res['id'] for res in search(['Doe'])]) \ == set([data.origin_url_1, data.origin_url_2]) assert [res['id'] for res in search(['John', 'Doe'])] \ == [data.origin_url_1] assert [res['id'] for res in search(['John'])] \ == [data.origin_url_1] assert not list(search(['John', 'Jane'])) def test_origin_intrinsic_metadata_search_fulltext_rank( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data # given tool_id = data.tools['swh-metadata-detector']['id'] # The following authors have "Random Person" to add some more content # to the JSON data, to work around normalization quirks when there # are few words (rank/(1+ln(nb_words)) is very sensitive to nb_words # for small values of nb_words). metadata1 = { 'author': [ 'Random Person', 'John Doe', 'Jane Doe', ] } metadata1_rev = { 'id': data.revision_id_1, 'metadata': metadata1, 'mappings': [], 'indexer_configuration_id': tool_id, } metadata1_origin = { 'id': data.origin_url_1, 'metadata': metadata1, 'mappings': [], 'indexer_configuration_id': tool_id, 'from_revision': data.revision_id_1, } metadata2 = { 'author': [ 'Random Person', 'Jane Doe', ] } metadata2_rev = { 'id': data.revision_id_2, 'metadata': metadata2, 'mappings': [], 'indexer_configuration_id': tool_id, } metadata2_origin = { 'id': data.origin_url_2, 'metadata': metadata2, 'mappings': [], 'indexer_configuration_id': tool_id, 'from_revision': data.revision_id_2, } # when storage.revision_intrinsic_metadata_add([metadata1_rev]) storage.origin_intrinsic_metadata_add([metadata1_origin]) storage.revision_intrinsic_metadata_add([metadata2_rev]) storage.origin_intrinsic_metadata_add([metadata2_origin]) # then search = storage.origin_intrinsic_metadata_search_fulltext assert [res['id'] for res in search(['Doe'])] \ == [data.origin_url_1, data.origin_url_2] assert [res['id'] for res in search(['Doe'], limit=1)] \ == [data.origin_url_1] assert [res['id'] for res in search(['John'])] \ == [data.origin_url_1] assert [res['id'] for res in search(['Jane'])] \ == [data.origin_url_2, data.origin_url_1] assert [res['id'] for res in search(['John', 'Jane'])] \ == [data.origin_url_1] def _fill_origin_intrinsic_metadata( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data tool1_id = data.tools['swh-metadata-detector']['id'] tool2_id = data.tools['swh-metadata-detector2']['id'] metadata1 = { '@context': 'foo', 'author': 'John Doe', } metadata1_rev = { 'id': data.revision_id_1, 'metadata': metadata1, 'mappings': ['npm'], 'indexer_configuration_id': tool1_id, } metadata1_origin = { 'id': data.origin_url_1, 'metadata': metadata1, 'mappings': ['npm'], 'indexer_configuration_id': tool1_id, 'from_revision': data.revision_id_1, } metadata2 = { '@context': 'foo', 'author': 'Jane Doe', } metadata2_rev = { 'id': data.revision_id_2, 'metadata': metadata2, 'mappings': ['npm', 'gemspec'], 'indexer_configuration_id': tool2_id, } metadata2_origin = { 'id': data.origin_url_2, 'metadata': metadata2, 'mappings': ['npm', 'gemspec'], 'indexer_configuration_id': tool2_id, 'from_revision': data.revision_id_2, } metadata3 = { '@context': 'foo', } metadata3_rev = { 'id': data.revision_id_3, 'metadata': metadata3, 'mappings': ['npm', 'gemspec'], 'indexer_configuration_id': tool2_id, } metadata3_origin = { 'id': data.origin_url_3, 'metadata': metadata3, 'mappings': ['pkg-info'], 'indexer_configuration_id': tool2_id, 'from_revision': data.revision_id_3, } storage.revision_intrinsic_metadata_add([metadata1_rev]) storage.origin_intrinsic_metadata_add([metadata1_origin]) storage.revision_intrinsic_metadata_add([metadata2_rev]) storage.origin_intrinsic_metadata_add([metadata2_origin]) storage.revision_intrinsic_metadata_add([metadata3_rev]) storage.origin_intrinsic_metadata_add([metadata3_origin]) def test_origin_intrinsic_metadata_search_by_producer( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data self._fill_origin_intrinsic_metadata( swh_indexer_storage_with_data) tool1 = data.tools['swh-metadata-detector'] tool2 = data.tools['swh-metadata-detector2'] endpoint = storage.origin_intrinsic_metadata_search_by_producer # test pagination # no 'page_token' param, return all origins result = endpoint(ids_only=True) assert result['origins'] \ == [data.origin_url_1, data.origin_url_2, data.origin_url_3] assert 'next_page_token' not in result # 'page_token' is < than origin_1, return everything result = endpoint(page_token=data.origin_url_1[:-1], ids_only=True) assert result['origins'] \ == [data.origin_url_1, data.origin_url_2, data.origin_url_3] assert 'next_page_token' not in result # 'page_token' is origin_3, return nothing result = endpoint(page_token=data.origin_url_3, ids_only=True) assert not result['origins'] assert 'next_page_token' not in result # test limit argument result = endpoint(page_token=data.origin_url_1[:-1], limit=2, ids_only=True) assert result['origins'] == [data.origin_url_1, data.origin_url_2] assert result['next_page_token'] == result['origins'][-1] result = endpoint(page_token=data.origin_url_1, limit=2, ids_only=True) assert result['origins'] == [data.origin_url_2, data.origin_url_3] assert 'next_page_token' not in result result = endpoint(page_token=data.origin_url_2, limit=2, ids_only=True) assert result['origins'] == [data.origin_url_3] assert 'next_page_token' not in result # test mappings filtering result = endpoint(mappings=['npm'], ids_only=True) assert result['origins'] == [data.origin_url_1, data.origin_url_2] assert 'next_page_token' not in result result = endpoint(mappings=['npm', 'gemspec'], ids_only=True) assert result['origins'] == [data.origin_url_1, data.origin_url_2] assert 'next_page_token' not in result result = endpoint(mappings=['gemspec'], ids_only=True) assert result['origins'] == [data.origin_url_2] assert 'next_page_token' not in result result = endpoint(mappings=['pkg-info'], ids_only=True) assert result['origins'] == [data.origin_url_3] assert 'next_page_token' not in result result = endpoint(mappings=['foobar'], ids_only=True) assert not result['origins'] assert 'next_page_token' not in result # test pagination + mappings result = endpoint(mappings=['npm'], limit=1, ids_only=True) assert result['origins'] == [data.origin_url_1] assert result['next_page_token'] == result['origins'][-1] # test tool filtering result = endpoint(tool_ids=[tool1['id']], ids_only=True) assert result['origins'] == [data.origin_url_1] assert 'next_page_token' not in result result = endpoint(tool_ids=[tool2['id']], ids_only=True) assert sorted(result['origins']) \ == [data.origin_url_2, data.origin_url_3] assert 'next_page_token' not in result result = endpoint(tool_ids=[tool1['id'], tool2['id']], ids_only=True) assert sorted(result['origins']) \ == [data.origin_url_1, data.origin_url_2, data.origin_url_3] assert 'next_page_token' not in result # test ids_only=False assert endpoint(mappings=['gemspec'])['origins'] \ == [{ 'id': data.origin_url_2, 'metadata': { '@context': 'foo', 'author': 'Jane Doe', }, 'mappings': ['npm', 'gemspec'], 'tool': tool2, 'from_revision': data.revision_id_2, }] def test_origin_intrinsic_metadata_stats( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data self._fill_origin_intrinsic_metadata( swh_indexer_storage_with_data) result = storage.origin_intrinsic_metadata_stats() assert result == { 'per_mapping': { 'gemspec': 1, 'npm': 2, 'pkg-info': 1, 'codemeta': 0, 'maven': 0, }, 'total': 3, 'non_empty': 2, } class TestIndexerStorageIndexerCondifuration: def test_indexer_configuration_add( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data tool = { 'tool_name': 'some-unknown-tool', 'tool_version': 'some-version', 'tool_configuration': {"debian-package": "some-package"}, } actual_tool = storage.indexer_configuration_get(tool) assert actual_tool is None # does not exist # add it actual_tools = list(storage.indexer_configuration_add([tool])) assert len(actual_tools) == 1 actual_tool = actual_tools[0] assert actual_tool is not None # now it exists new_id = actual_tool.pop('id') assert actual_tool == tool actual_tools2 = list(storage.indexer_configuration_add([tool])) actual_tool2 = actual_tools2[0] assert actual_tool2 is not None # now it exists new_id2 = actual_tool2.pop('id') assert new_id == new_id2 assert actual_tool == actual_tool2 def test_indexer_configuration_add_multiple( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data tool = { 'tool_name': 'some-unknown-tool', 'tool_version': 'some-version', 'tool_configuration': {"debian-package": "some-package"}, } actual_tools = list(storage.indexer_configuration_add([tool])) assert len(actual_tools) == 1 new_tools = [tool, { 'tool_name': 'yet-another-tool', 'tool_version': 'version', 'tool_configuration': {}, }] actual_tools = list(storage.indexer_configuration_add(new_tools)) assert len(actual_tools) == 2 # order not guaranteed, so we iterate over results to check for tool in actual_tools: _id = tool.pop('id') assert _id is not None assert tool in new_tools def test_indexer_configuration_get_missing( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data tool = { 'tool_name': 'unknown-tool', 'tool_version': '3.1.0rc2-31-ga2cbb8c', 'tool_configuration': {"command_line": "nomossa "}, } actual_tool = storage.indexer_configuration_get(tool) assert actual_tool is None def test_indexer_configuration_get( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data tool = { 'tool_name': 'nomos', 'tool_version': '3.1.0rc2-31-ga2cbb8c', 'tool_configuration': {"command_line": "nomossa "}, } actual_tool = storage.indexer_configuration_get(tool) assert actual_tool expected_tool = tool.copy() del actual_tool['id'] assert expected_tool == actual_tool def test_indexer_configuration_metadata_get_missing_context( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data tool = { 'tool_name': 'swh-metadata-translator', 'tool_version': '0.0.1', 'tool_configuration': {"context": "unknown-context"}, } actual_tool = storage.indexer_configuration_get(tool) assert actual_tool is None def test_indexer_configuration_metadata_get( self, swh_indexer_storage_with_data): storage, data = swh_indexer_storage_with_data tool = { 'tool_name': 'swh-metadata-translator', 'tool_version': '0.0.1', 'tool_configuration': {"type": "local", "context": "NpmMapping"}, } storage.indexer_configuration_add([tool]) actual_tool = storage.indexer_configuration_get(tool) assert actual_tool expected_tool = tool.copy() expected_tool['id'] = actual_tool['id'] assert expected_tool == actual_tool class TestIndexerStorageMisc: """Misc endpoints tests for the IndexerStorage. """ def test_check_config(self, swh_indexer_storage): storage = swh_indexer_storage assert storage.check_config(check_write=True) assert storage.check_config(check_write=False) def test_types(self, swh_indexer_storage): """Checks all methods of StorageInterface are implemented by this backend, and that they have the same signature.""" # Create an instance of the protocol (which cannot be instantiated # directly, so this creates a subclass, then instantiates it) interface = type('_', (IndexerStorageInterface,), {})() assert 'content_mimetype_add' in dir(interface) missing_methods = [] for meth_name in dir(interface): if meth_name.startswith('_'): continue interface_meth = getattr(interface, meth_name) try: concrete_meth = getattr(swh_indexer_storage, meth_name) except AttributeError: missing_methods.append(meth_name) continue expected_signature = inspect.signature(interface_meth) actual_signature = inspect.signature(concrete_meth) assert expected_signature == actual_signature, meth_name assert missing_methods == []