path: root/wqflask/maintenance/quick_search_table.py

"""Creates a table used for the quick search feature.

One column contains the terms to match the user's search against. Another contains the result
fields in json format

Results will be returned for each of several trait types: mRNA assays, phenotypes, genotypes, and
(maybe later) genes

For each trait type, the results for each species should be given This table will then list
each trait, its dataset, and several columns determined by its trait type (phenotype, genotype, etc)

"""

from __future__ import absolute_import, division, print_function

#import sys
#sys.path.append("../../..")

import simplejson as json

import sqlalchemy as sa
from sqlalchemy.orm import scoped_session, sessionmaker
from sqlalchemy.ext.declarative import declarative_base

#from pprint import pformat as pf

import zach_settings as settings

Engine = sa.create_engine(settings.SQLALCHEMY_DATABASE_URI,
                       #encoding='utf-8',
                       #client_encoding='utf-8',
                       #echo="debug",w
                       )

Session = scoped_session(sessionmaker(bind=Engine)) #, extension=VersionedListener()))

Base = declarative_base(bind=Engine)
Metadata = sa.MetaData()
Metadata.bind = Engine

class PublishXRef(Base):
    """Class that corresponds with the PublishXRef table in the database.
    
    The PublishXRef table links phenotype traits and their publications.
    
    This class is used to add phenotype traits to the quick search table.
    
    """
    
    __tablename__ = 'PublishXRef'

    Id = sa.Column(sa.Integer, primary_key=True)
    InbredSetId = sa.Column(sa.Integer, primary_key=True)
    PhenotypeId = sa.Column(sa.Integer)
    PublicationId = sa.Column(sa.Integer)
    DataId = sa.Column(sa.Integer)
    Locus = sa.Column(sa.Text)
    LRS = sa.Column(sa.Float)
    additive = sa.Column(sa.Float)
    Sequence = sa.Column(sa.Integer)
    comments = sa.Column(sa.Text)

    @classmethod
    def run(cls):
        """Connects to database and inserts phenotype trait info into the Quicksearch table."""
        conn = Engine.connect()
        counter = 0
        for pub_row in page_query(Session.query(cls)):   #all()
            values = {}
            values['table_name'] = cls.__tablename__
            values['the_key'] = json.dumps([pub_row.Id, pub_row.InbredSetId])
            values['terms'] = cls.get_unique_terms(pub_row.Id, pub_row.InbredSetId)
            print("terms is:", values['terms'])
            if values['terms']:
                values['result_fields'] = cls.get_result_fields(pub_row.Id, pub_row.InbredSetId)
                ins = QuickSearch.insert().values(**values)
                conn.execute(ins)
            counter += 1
            print("Done:", counter)

    @staticmethod
    def get_unique_terms(publishxref_id, inbredset_id):
        """Finds unique terms for each item in the PublishXRef table to match a query against"""
        results = Session.query(
                "pre_publication_description",
                "post_publication_description",
                "pre_publication_abbreviation",
                "post_publication_abbreviation",
                "publication_title"
            ).from_statement(
                "SELECT Phenotype.Pre_publication_description as pre_publication_description, "
                "Phenotype.Post_publication_description as post_publication_description, "
                "Phenotype.Pre_publication_abbreviation as pre_publication_abbreviation, "
                "Phenotype.Post_publication_abbreviation as post_publication_abbreviation, "
                "Publication.Title as publication_title "
                "FROM Phenotype, Publication, PublishXRef "
                "WHERE PublishXRef.Id = :publishxref_id and "
                "PublishXRef.InbredSetId = :inbredset_id and "
                "PublishXRef.PhenotypeId = Phenotype.Id and "
                "PublishXRef.PublicationId = Publication.Id ").params(publishxref_id=publishxref_id,
                                                            inbredset_id=inbredset_id).all()

        unique = set()
        print("results: ", results)
        if len(results):
            for item in results[0]:
                #print("locals:", locals())
                if not item:
                    continue
                for token in item.split():
                    if token.startswith(('(','[')):
                        token = token[1:]
                    if token.endswith((')', ']')):
                        token = token[:-1]
                    if token.endswith(';'):
                        token = token[:-1]
                    if len(token) > 2:
                        try:
                            # This hopefully ensures that the token is utf-8
                            token = token.encode('utf-8')
                            print(" ->", token)
                        except UnicodeDecodeError:
                            print("\n-- UDE \n")
                            # Can't get it into utf-8, we won't use it
                            continue 
    
                        unique.add(token)
        #print("\nUnique terms are: {}\n".format(unique))
        return " ".join(unique)            

    @staticmethod
    def get_result_fields(publishxref_id, inbredset_id):
        """Gets the result fields (columns) that appear on the result page as a json string"""
        results = Session.query(
                "phenotype_id",
                "species",
                "group_name",
                "description",
                "lrs",
                "publication_id",
                "pubmed_id",
                "year",
                "authors"
            ).from_statement(
                "SELECT PublishXRef.PhenotypeId as phenotype_id, "
                "Species.Name as species, "
                "InbredSet.Name as group_name, "
                "Phenotype.Original_description as description, "
                "PublishXRef.LRS as lrs, "
                "PublishXRef.PublicationId as publication_id, "
                "Publication.PubMed_ID as pubmed_id, "
                "Publication.Year as year, "
                "Publication.Authors as authors "
                "FROM PublishXRef, "
                "Phenotype, "
                "Publication, "
                "InbredSet, "
                "Species "
                "WHERE PublishXRef.Id = :publishxref_id and "
                "PublishXRef.InbredSetId = :inbredset_id and "
                "PublishXRef.PhenotypeId = Phenotype.Id and "
                "PublishXRef.PublicationId = Publication.Id and "
                "InbredSet.Id = :inbredset_id and "
                "Species.Id = InbredSet.SpeciesId ").params(publishxref_id=publishxref_id,
                                                            inbredset_id=inbredset_id).all()

        assert len(set(result for result in results)) == 1, "Different results or no results"

        result = results[0]
        result = row2dict(result)
        try:
            json_results = json.dumps(result, sort_keys=True)
        except UnicodeDecodeError:
            print("\n\nTrying to massage unicode\n\n")
            for key, value in result.iteritems():
                print("\tkey is:", key)
                print("\tvalue is:", value)
                if isinstance(value, basestring):
                    result[key] = value.decode('utf-8', errors='ignore')
            json_results = json.dumps(result, sort_keys=True)

        return json_results


class GenoXRef(Base):
    """Class that corresponds with the GenoXRef table in the database.
    
    The GenoXRef table links genotype traits and their data.
    
    This class is used to add genotype traits to the quick search table.
    
    """
    
    __tablename__ = 'GenoXRef'
    
    GenoFreezeId = sa.Column(sa.Integer, primary_key=True)
    GenoId = sa.Column(sa.Integer, primary_key=True)
    DataId = sa.Column(sa.Integer)
    cM = sa.Column(sa.Float)
    Used_for_mapping = sa.Column(sa.Text)

    @classmethod
    def run(cls):
        """Connects to database and inserts genotype trait info into the Quicksearch table."""
        conn = Engine.connect()
        counter = 0
        for item in page_query(Session.query(cls)):   #all()
            values = {}
            values['table_name'] = cls.__tablename__
            values['the_key'] = json.dumps([item.GenoId, item.GenoFreezeId])
            values['terms'] = cls.get_unique_terms(item.GenoId)
            print("terms is:", values['terms'])
            if values['terms']:
                values['result_fields'] = cls.get_result_fields(item.GenoId, item.GenoFreezeId)
                ins = QuickSearch.insert().values(**values)
                conn.execute(ins)
            counter += 1
            print("Done:", counter)

    @staticmethod
    def get_unique_terms(geno_id):
        """Finds unique terms for each item in the GenoXRef table to match a query against"""
        print("geno_id: ", geno_id)
        results = Session.query(
                "name",
                "marker_name"
            ).from_statement(
                "SELECT Geno.Name as name, "
                "Geno.Marker_Name as marker_name "
                "FROM Geno "
                "WHERE Geno.Id = :geno_id ").params(geno_id=geno_id).all()
        
        unique = set()
        if len(results):
            for item in results[0]:
                #print("locals:", locals())
                if not item:
                    continue
                for token in item.split():
                    if len(token) > 2:
                        try:
                            # This hopefully ensures that the token is utf-8
                            token = token.encode('utf-8')
                            print(" ->", token)
                        except UnicodeDecodeError:
                            print("\n-- UDE \n")
                            # Can't get it into utf-8, we won't use it
                            continue 
                        
                        unique.add(token)
        return " ".join(unique)


    @staticmethod
    def get_result_fields(geno_id, dataset_id):
        """Gets the result fields (columns) that appear on the result page as a json string"""
        results = Session.query(
                "name",
                "marker_name",
                "group_name",
                "species",
                "dataset",
                "dataset_name",
                "chr", "mb",
                "source"
            ).from_statement(
                "SELECT Geno.Name as name, "
                "Geno.Marker_Name as marker_name, "
                "InbredSet.Name as group_name, "
                "Species.Name as species, "
                "GenoFreeze.Name as dataset, "
                "GenoFreeze.FullName as dataset_name, "
                "Geno.Chr as chr, "
                "Geno.Mb as mb, "
                "Geno.Source as source "
                "FROM Geno, "
                "GenoXRef, "
                "GenoFreeze, "
                "InbredSet, "
                "Species "
                "WHERE Geno.Id = :geno_id and "
                "GenoXRef.GenoId = Geno.Id and "
                "GenoFreeze.Id = :dataset_id and "
                "GenoXRef.GenoFreezeId = GenoFreeze.Id and "
                "InbredSet.Id = GenoFreeze.InbredSetId and "
                "InbredSet.SpeciesId = Species.Id ").params(geno_id=geno_id,
                                                                    dataset_id=dataset_id).all()
        assert len(set(result for result in results)) == 1, "Different results"
        
        result = results[0]
        result = row2dict(result)
        try:
            json_results = json.dumps(result, sort_keys=True)
        except UnicodeDecodeError:
            print("\n\nTrying to massage unicode\n\n")
            for key, value in result.iteritems():
                print("\tkey is:", key)
                print("\tvalue is:", value)
                if isinstance(value, basestring):
                    result[key] = value.decode('utf-8', errors='ignore')
            json_results = json.dumps(result, sort_keys=True)

        return json_results    

class ProbeSetXRef(Base):
    """Class that corresponds with the ProbeSetXRef table in the database.
    
    The ProbeSetXRef table links mRNA expression traits and their sample data.
    
    This class is used to add mRNA expression traits to the quick search table.
    
    """
    
    __tablename__ = 'ProbeSetXRef'
    
    ProbeSetFreezeId = sa.Column(sa.Integer, primary_key=True)
    ProbeSetId = sa.Column(sa.Integer, primary_key=True)
    DataId = sa.Column(sa.Integer, unique=True)
    Locus_old = sa.Column(sa.Text)
    LRS_old = sa.Column(sa.Float)
    pValue_old = sa.Column(sa.Float)
    mean = sa.Column(sa.Float)
    se = sa.Column(sa.Float)
    Locus = sa.Column(sa.Text)
    LRS = sa.Column(sa.Float)
    pValue = sa.Column(sa.Float)
    additive = sa.Column(sa.Float)
    h2 = sa.Column(sa.Float)

    @classmethod
    def run(cls):
        """Connects to db and inserts mRNA expression trait info into the Quicksearch table."""
        conn = Engine.connect()
        counter = 0
        for ps_row in page_query(Session.query(cls)):   #all()
            values = {}
            values['table_name'] = cls.__tablename__
            values['the_key'] = json.dumps([ps_row.ProbeSetId, ps_row.ProbeSetFreezeId])
            values['terms'] = cls.get_unique_terms(ps_row.ProbeSetId)
            print("terms is:", values['terms'])
            values['result_fields'] = cls.get_result_fields(ps_row.ProbeSetId,
                                                            ps_row.ProbeSetFreezeId)
            if values['result_fields'] == None:
                continue
            ins = QuickSearch.insert().values(**values)
            conn.execute(ins)
            counter += 1
            print("Done:", counter)
    
    @staticmethod
    def get_unique_terms(probeset_id):
        """Finds unique terms for each item in the ProbeSetXRef table to match a query against"""
        results = Session.query(
                "name",
                "symbol",
                "description",
                "alias"
            ).from_statement(
                "SELECT ProbeSet.Name as name, "
                "ProbeSet.Symbol as symbol, "
                "ProbeSet.description as description, "
                "ProbeSet.alias as alias "
                "FROM ProbeSet "
                "WHERE ProbeSet.Id = :probeset_id ").params(probeset_id=probeset_id).all()

        unique = set()
        if len(results):
            for item in results[0]:
                if not item:
                    continue
                for token in item.split():
                    if token.startswith(('(','[')):
                        token = token[1:]
                    if token.endswith((')', ']')):
                        token = token[:-1]
                    if token.endswith(';'):
                        token = token[:-1]
                    if len(token) > 2:
                        try:
                            # This hopefully ensures that the token is utf-8
                            token = token.encode('utf-8')
                            print(" ->", token)
                        except UnicodeDecodeError:
                            print("\n-- UDE \n")
                            # Can't get it into utf-8, we won't use it
                            continue 
                        
                        unique.add(token)
            return " ".join(unique)


    @staticmethod
    def get_result_fields(probeset_id, dataset_id):
        """Gets the result fields (columns) that appear on the result page as a json string"""
        print("probeset_id: ", probeset_id)
        print("dataset_id: ", dataset_id)
        results = Session.query(
                "name",
                "species",
                "group_name",
                "dataset",
                "dataset_name",
                "symbol",
                "description",
                "chr", "mb",
                "lrs", "mean",
                "genbank_id",
                "gene_id",
                "chip_id",
                "chip_name"
            ).from_statement(
                "SELECT ProbeSet.Name as name, "
                "Species.Name as species, "
                "InbredSet.Name as group_name, "
                "ProbeSetFreeze.Name as dataset, "
                "ProbeSetFreeze.FullName as dataset_name, "
                "ProbeSet.Symbol as symbol, "
                "ProbeSet.description as description, "
                "ProbeSet.Chr as chr, "
                "ProbeSet.Mb as mb, "
                "ProbeSetXRef.LRS as lrs, "
                "ProbeSetXRef.mean as mean, "
                "ProbeSet.GenbankId as genbank_id, "
                "ProbeSet.GeneId as gene_id, "
                "ProbeSet.ChipId as chip_id, "
                "GeneChip.Name as chip_name "
                "FROM ProbeSet, "
                "ProbeSetXRef, "
                "ProbeSetFreeze, "
                "ProbeFreeze, "
                "InbredSet, "
                "Species, "
                "GeneChip "
                "WHERE ProbeSetXRef.ProbeSetId = :probeset_id and "
                "ProbeSetXRef.ProbeSetFreezeId = :dataset_id and "
                "ProbeSetXRef.ProbeSetId = ProbeSet.Id and "
                "ProbeSet.ChipId = GeneChip.Id and "
                "ProbeSetXRef.ProbeSetFreezeId = ProbeSetFreeze.Id and "
                "ProbeSetFreeze.ProbeFreezeId = ProbeFreeze.Id and "
                "ProbeFreeze.InbredSetId = InbredSet.Id and "
                "InbredSet.SpeciesId = Species.Id ").params(probeset_id=probeset_id,
                                                                    dataset_id=dataset_id).all()
            
        if len(set(result for result in results)) != 1:
            return None

        result = results[0]
        result = row2dict(result)
        try:
            json_results = json.dumps(result, sort_keys=True)
        except UnicodeDecodeError:
            print("\n\nTrying to massage unicode\n\n")
            for key, value in result.iteritems():
                print("\tkey is:", key)
                print("\tvalue is:", value)
                if isinstance(value, basestring):
                    result[key] = value.decode('utf-8', errors='ignore')
            json_results = json.dumps(result, sort_keys=True)

        return json_results    

QuickSearch = sa.Table("QuickSearch", Metadata,
        # table_name is the table that item is inserted from
        sa.Column('table_name', sa.String(15),
                  primary_key=True, nullable=False, autoincrement=False), 
        sa.Column('the_key', sa.String(30),
                  primary_key=True, nullable=False, autoincrement=False), # key in database table
        sa.Column('terms', sa.Text), # terms to compare search string with
        sa.Column('result_fields', sa.Text),  # json
        mysql_engine = 'MyISAM',
                    )

QuickSearch.drop(Engine, checkfirst=True)
Metadata.create_all(Engine)


def row2dict(row):
    """From http://stackoverflow.com/a/2848519/1175849"""
    return dict(zip(row.keys(), row))


def page_query(query):
    """From http://stackoverflow.com/a/1217947/1175849"""
    offset = 0
    while True:
        rrr = False
        for elem in query.limit(1000).offset(offset):
            rrr = True
            yield elem
        offset += 1000
        if not rrr:
            break


def main():
    """Populate the QuickSearch table that is used with the quick search features.
    
    Add all items from the ProbeSetXRef, GenoXRef, and PublishXRef tables to the QuickSearch tables.
    
    """

    GenoXRef.run()
    PublishXRef.run()
    ProbeSetXRef.run()

if __name__ == "__main__":
    main()