aboutsummaryrefslogtreecommitdiff
path: root/wqflask
diff options
context:
space:
mode:
authorArun Isaac2022-09-21 13:41:15 +0530
committerArun Isaac2022-09-26 13:39:45 +0530
commit73327a58707d7321e2d4c616da442e598ed13c57 (patch)
tree5c138264bd4ee5bf5d210dbbaefcb449dad312ab /wqflask
parent66a6593daf4580f3056c96ccfc658b556ea619fc (diff)
downloadgenenetwork2-73327a58707d7321e2d4c616da442e598ed13c57.tar.gz
Add indexing script for global search.
* wqflask/scripts/index.py: New file.
Diffstat (limited to 'wqflask')
-rw-r--r--wqflask/scripts/index.py147
1 files changed, 147 insertions, 0 deletions
diff --git a/wqflask/scripts/index.py b/wqflask/scripts/index.py
new file mode 100644
index 00000000..fe2b819c
--- /dev/null
+++ b/wqflask/scripts/index.py
@@ -0,0 +1,147 @@
+"""This script must be run each time the database is updated. It runs
+queries against the SQL database, indexes the results and builds a
+xapian index. This xapian index is later used in providing search
+through the web interface.
+
+"""
+
+from functools import partial
+import json
+import xapian
+
+from utility.monads import MonadicDictCursor
+from wqflask.database import database_connection, xapian_writable_database
+
+
+def index_text(termgenerator, text):
+ """Index text and increase term position."""
+ termgenerator.index_text(text)
+ termgenerator.increase_termpos()
+
+
+# pylint: disable=missing-function-docstring
+def main():
+ with database_connection() as conn, conn.cursor(MonadicDictCursor) as cursor:
+ # FIXME: Some Max LRS values in the DB are wrongly listed as
+ # 0.000, but shouldn't be displayed. Make them NULLs in the
+ # database.
+ cursor.execute("""
+ SELECT ProbeSet.Name AS name,
+ ProbeSet.Symbol AS symbol,
+ ProbeSet.description AS description,
+ ProbeSet.Chr AS chr,
+ ProbeSet.Mb AS mb,
+ ProbeSet.alias AS alias,
+ ProbeSet.GenbankId AS genbankid,
+ ProbeSet.UniGeneId AS unigeneid,
+ ProbeSet.Probe_Target_Description AS probe_target_description,
+ ProbeSetFreeze.Name AS dataset,
+ ProbeSetFreeze.FullName AS dataset_fullname,
+ ProbeSetFreeze.Id AS dataset_id,
+ Species.Name AS species,
+ InbredSet.Name AS `group`,
+ Tissue.Name AS tissue,
+ ProbeSetXRef.Mean AS mean,
+ ProbeSetXRef.LRS AS lrs,
+ ProbeSetXRef.additive AS additive,
+ Geno.Chr as geno_chr,
+ Geno.Mb as geno_mb
+ FROM Species
+ INNER JOIN InbredSet ON InbredSet.SpeciesId = Species.Id
+ INNER JOIN ProbeFreeze ON ProbeFreeze.InbredSetId = InbredSet.Id
+ INNER JOIN Tissue ON ProbeFreeze.TissueId = Tissue.Id
+ INNER JOIN ProbeSetFreeze ON ProbeSetFreeze.ProbeFreezeId = ProbeFreeze.Id
+ INNER JOIN ProbeSetXRef ON ProbeSetXRef.ProbeSetFreezeId = ProbeSetFreeze.Id
+ INNER JOIN ProbeSet ON ProbeSet.Id = ProbeSetXRef.ProbeSetId
+ LEFT JOIN Geno ON ProbeSetXRef.Locus = Geno.Name AND Geno.SpeciesId = Species.Id
+ WHERE ProbeSetFreeze.confidentiality < 1 AND ProbeSetFreeze.public > 0
+ """)
+ termgenerator = xapian.TermGenerator()
+ termgenerator.set_stemmer(xapian.Stem("en"))
+ indexer = partial(index_text, termgenerator)
+
+ # pylint: disable=invalid-name
+ with xapian_writable_database() as db:
+ for trait in cursor.fetchall():
+ doc = xapian.Document()
+ termgenerator.set_document(doc)
+
+ # Index text.
+ trait["name"].bind(indexer)
+ trait["description"].bind(indexer)
+ trait["symbol"].bind(indexer)
+ trait.pop("alias").bind(indexer)
+ trait.pop("genbankid").bind(indexer)
+ trait.pop("unigeneid").bind(indexer)
+ trait.pop("probe_target_description").bind(indexer)
+
+ # Identify document as type "gene". We use the XT
+ # prefix to indicate the type.
+ doc.add_boolean_term("XTgene")
+ doc.set_data(json.dumps(trait.data))
+ # Write document into xapian database.
+ idterm = trait["name"].bind(lambda name: "Q" + name)
+ doc.add_boolean_term(idterm)
+ db.replace_document(idterm, doc)
+
+ cursor.execute("""
+ SELECT Species.Name AS species,
+ InbredSet.Name AS `group`,
+ PublishFreeze.Name AS dataset,
+ PublishFreeze.FullName AS dataset_fullname,
+ PublishXRef.Id AS name,
+ COALESCE(Phenotype.Post_publication_abbreviation, Phenotype.Pre_publication_abbreviation) AS abbreviation,
+ COALESCE(Phenotype.Post_publication_description, Phenotype.Pre_publication_description) AS description,
+ Phenotype.Lab_code,
+ Publication.Abstract,
+ Publication.Title,
+ Publication.Authors AS authors,
+ Publication.Year AS year,
+ Publication.PubMed_ID AS pubmed_id,
+ PublishXRef.LRS as lrs,
+ PublishXRef.additive,
+ InbredSet.InbredSetCode AS inbredsetcode,
+ PublishXRef.mean,
+ PublishFreeze.Id AS dataset_id,
+ Geno.Chr as geno_chr,
+ Geno.Mb as geno_mb
+ FROM Species
+ INNER JOIN InbredSet ON InbredSet.SpeciesId = Species.Id
+ INNER JOIN PublishFreeze ON PublishFreeze.InbredSetId = InbredSet.Id
+ INNER JOIN PublishXRef ON PublishXRef.InbredSetId = InbredSet.Id
+ INNER JOIN Phenotype ON PublishXRef.PhenotypeId = Phenotype.Id
+ INNER JOIN Publication ON PublishXRef.PublicationId = Publication.Id
+ LEFT JOIN Geno ON PublishXRef.Locus = Geno.Name AND Geno.SpeciesId = Species.Id
+ """)
+ with xapian_writable_database() as db:
+ for i, trait in enumerate(cursor.fetchall()):
+ doc = xapian.Document()
+ termgenerator.set_document(doc)
+
+ # Index text.
+ trait.pop("abbreviation").bind(indexer)
+ trait["description"].bind(indexer)
+ trait.pop("Lab_code").bind(indexer)
+ trait.pop("Abstract").bind(indexer)
+ trait.pop("Title").bind(indexer)
+ trait["authors"].bind(indexer)
+ trait["inbredsetcode"].bind(indexer)
+
+ # Convert name from integer to string.
+ trait["name"] = trait["name"].map(str)
+ # Split comma-separated authors into a list.
+ trait["authors"] = trait["authors"].map(
+ lambda s: [author.strip() for author in s.split(",")])
+
+ # Identify document as type "phenotype". We use the XT
+ # prefix to indicate the type.
+ doc.add_boolean_term("XTphenotype")
+ # Write document into xapian database.
+ doc.set_data(json.dumps(trait.data))
+ idterm = f"Q{i}"
+ doc.add_boolean_term(idterm)
+ db.replace_document(idterm, doc)
+
+
+if __name__ == "__main__":
+ main()