diff options
Diffstat (limited to 'gn3')
| -rw-r--r-- | gn3/computations/heatmap.py | 177 | ||||
| -rw-r--r-- | gn3/computations/slink.py | 2 | ||||
| -rw-r--r-- | gn3/db/datasets.py | 291 | ||||
| -rw-r--r-- | gn3/db/traits.py | 628 | ||||
| -rw-r--r-- | gn3/function_helpers.py | 36 |
5 files changed, 1036 insertions, 98 deletions
diff --git a/gn3/computations/heatmap.py b/gn3/computations/heatmap.py new file mode 100644 index 0000000..a0e778a --- /dev/null +++ b/gn3/computations/heatmap.py @@ -0,0 +1,177 @@ +""" +This module will contain functions to be used in computation of the data used to +generate various kinds of heatmaps. +""" + +from functools import reduce +from typing import Any, Dict, Sequence +from gn3.computations.slink import slink +from gn3.db.traits import retrieve_trait_data, retrieve_trait_info +from gn3.computations.correlations2 import compute_correlation + +def export_trait_data( + trait_data: dict, strainlist: Sequence[str], dtype: str = "val", + var_exists: bool = False, n_exists: bool = False): + """ + Export data according to `strainlist`. Mostly used in calculating + correlations. + + DESCRIPTION: + Migrated from + https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/base/webqtlTrait.py#L166-L211 + + PARAMETERS + trait: (dict) + The dictionary of key-value pairs representing a trait + strainlist: (list) + A list of strain names + type: (str) + ... verify what this is ... + var_exists: (bool) + A flag indicating existence of variance + n_exists: (bool) + A flag indicating existence of ndata + """ + def __export_all_types(tdata, strain): + sample_data = [] + if tdata[strain]["val"]: + sample_data.append(tdata[strain]["val"]) + if var_exists: + if tdata[strain].var: + sample_data.append(tdata[strain]["var"]) + else: + sample_data.append(None) + if n_exists: + if tdata[strain]["ndata"]: + sample_data.append(tdata[strain]["ndata"]) + else: + sample_data.append(None) + else: + if var_exists and n_exists: + sample_data += [None, None, None] + elif var_exists or n_exists: + sample_data += [None, None] + else: + sample_data.append(None) + + return tuple(sample_data) + + def __exporter(accumulator, strain): + # pylint: disable=[R0911] + if trait_data.has_key(strain): + if dtype == "val": + return accumulator + (trait_data[strain]["val"], ) + if dtype == "var": + return accumulator + (trait_data[strain]["var"], ) + if dtype == "N": + return trait_data[strain]["ndata"] + if dtype == "all": + return accumulator + __export_all_types(trait_data, strain) + raise KeyError("Type `%s` is incorrect" % dtype) + if var_exists and n_exists: + return accumulator + (None, None, None) + if var_exists or n_exists: + return accumulator + (None, None) + return accumulator + (None,) + + return reduce(__exporter, strainlist, tuple()) + +def trait_display_name(trait: Dict): + """ + Given a trait, return a name to use to display the trait on a heatmap. + + DESCRIPTION + Migrated from + https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/base/webqtlTrait.py#L141-L157 + """ + if trait.get("db", None) and trait.get("trait_name", None): + if trait["db"]["dataset_type"] == "Temp": + desc = trait["description"] + if desc.find("PCA") >= 0: + return "%s::%s" % ( + trait["db"]["displayname"], + desc[desc.rindex(':')+1:].strip()) + return "%s::%s" % ( + trait["db"]["displayname"], + desc[:desc.index('entered')].strip()) + prefix = "%s::%s" % ( + trait["db"]["dataset_name"], trait["trait_name"]) + if trait["cellid"]: + return "%s::%s" % (prefix, trait["cellid"]) + return prefix + return trait["description"] + +def cluster_traits(traits_data_list: Sequence[Dict]): + """ + Clusters the trait values. + + DESCRIPTION + Attempts to replicate the clustering of the traits, as done at + https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/heatmap/Heatmap.py#L138-L162 + """ + def __compute_corr(tdata_i, tdata_j): + if tdata_j[0] < tdata_i[0]: + corr_vals = compute_correlation(tdata_i, tdata_j) + corr = corr_vals[0] + if (1 - corr) < 0: + return 0.0 + return 1 - corr + return 0.0 + + def __cluster(tdata_i): + return tuple( + __compute_corr(tdata_i, tdata_j) + for tdata_j in enumerate(traits_data_list)) + + return tuple(__cluster(tdata_i) for tdata_i in enumerate(traits_data_list)) + +def heatmap_data(formd, search_result, conn: Any): + """ + heatmap function + + DESCRIPTION + This function is an attempt to reproduce the initialisation at + https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/heatmap/Heatmap.py#L46-L64 + and also the clustering and slink computations at + https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/heatmap/Heatmap.py#L138-L165 + with the help of the `gn3.computations.heatmap.cluster_traits` function. + + It does not try to actually draw the heatmap image. + + PARAMETERS: + TODO: Elaborate on the parameters here... + """ + threshold = 0 # webqtlConfig.PUBLICTHRESH + cluster_checked = formd.formdata.getvalue("clusterCheck", "") + strainlist = [ + strain for strain in formd.strainlist if strain not in formd.parlist] + genotype = formd.genotype + + def __retrieve_traitlist_and_datalist(threshold, fullname): + trait = retrieve_trait_info(threshold, fullname, conn) + return ( + trait, + export_trait_data(retrieve_trait_data(trait, conn), strainlist)) + + traits_details = [ + __retrieve_traitlist_and_datalist(threshold, fullname) + for fullname in search_result] + traits_list = map(lambda x: x[0], traits_details) + traits_data_list = map(lambda x: x[1], traits_details) + + return { + "target_description_checked": formd.formdata.getvalue( + "targetDescriptionCheck", ""), + "cluster_checked": cluster_checked, + "slink_data": ( + slink(cluster_traits(traits_data_list)) + if cluster_checked else False), + "sessionfile": formd.formdata.getvalue("session"), + "genotype": genotype, + "nLoci": sum(map(len, genotype)), + "strainlist": strainlist, + "ppolar": formd.ppolar, + "mpolar":formd.mpolar, + "traits_list": traits_list, + "traits_data_list": traits_data_list + } diff --git a/gn3/computations/slink.py b/gn3/computations/slink.py index 23d3d88..5953e6b 100644 --- a/gn3/computations/slink.py +++ b/gn3/computations/slink.py @@ -7,7 +7,7 @@ slink: TODO: Describe what the function does... """ import logging -from typing import List, Tuple, Union, Sequence +from typing import Union, Sequence NumType = Union[int, float] SeqOfNums = Sequence[NumType] diff --git a/gn3/db/datasets.py b/gn3/db/datasets.py new file mode 100644 index 0000000..53d6811 --- /dev/null +++ b/gn3/db/datasets.py @@ -0,0 +1,291 @@ +""" +This module contains functions relating to specific trait dataset manipulation +""" +from typing import Any + +def retrieve_probeset_trait_dataset_name( + threshold: int, name: str, connection: Any): + """ + Get the ID, DataScale and various name formats for a `ProbeSet` trait. + """ + query = ( + "SELECT Id, Name, FullName, ShortName, DataScale " + "FROM ProbeSetFreeze " + "WHERE " + "public > %(threshold)s " + "AND " + "(Name = %(name)s OR FullName = %(name)s OR ShortName = %(name)s)") + with connection.cursor() as cursor: + cursor.execute( + query, + { + "threshold": threshold, + "name": name + }) + return dict(zip( + ["dataset_id", "dataset_name", "dataset_fullname", + "dataset_shortname", "dataset_datascale"], + cursor.fetchone)) + +def retrieve_publish_trait_dataset_name( + threshold: int, name: str, connection: Any): + """ + Get the ID, DataScale and various name formats for a `Publish` trait. + """ + query = ( + "SELECT Id, Name, FullName, ShortName " + "FROM PublishFreeze " + "WHERE " + "public > %(threshold)s " + "AND " + "(Name = %(name)s OR FullName = %(name)s OR ShortName = %(name)s)") + with connection.cursor() as cursor: + cursor.execute( + query, + { + "threshold": threshold, + "name": name + }) + return dict(zip( + ["dataset_id", "dataset_name", "dataset_fullname", + "dataset_shortname"], + cursor.fetchone)) + +def retrieve_geno_trait_dataset_name( + threshold: int, name: str, connection: Any): + """ + Get the ID, DataScale and various name formats for a `Geno` trait. + """ + query = ( + "SELECT Id, Name, FullName, ShortName " + "FROM GenoFreeze " + "WHERE " + "public > %(threshold)s " + "AND " + "(Name = %(name)s OR FullName = %(name)s OR ShortName = %(name)s)") + with connection.cursor() as cursor: + cursor.execute( + query, + { + "threshold": threshold, + "name": name + }) + return dict(zip( + ["dataset_id", "dataset_name", "dataset_fullname", + "dataset_shortname"], + cursor.fetchone)) + +def retrieve_temp_trait_dataset_name( + threshold: int, name: str, connection: Any): + """ + Get the ID, DataScale and various name formats for a `Temp` trait. + """ + query = ( + "SELECT Id, Name, FullName, ShortName " + "FROM TempFreeze " + "WHERE " + "public > %(threshold)s " + "AND " + "(Name = %(name)s OR FullName = %(name)s OR ShortName = %(name)s)") + with connection.cursor() as cursor: + cursor.execute( + query, + { + "threshold": threshold, + "name": name + }) + return dict(zip( + ["dataset_id", "dataset_name", "dataset_fullname", + "dataset_shortname"], + cursor.fetchone)) + +def retrieve_dataset_name( + trait_type: str, threshold: int, trait_name: str, dataset_name: str, + conn: Any): + """ + Retrieve the name of a trait given the trait's name + + This is extracted from the `webqtlDataset.retrieveName` function as is + implemented at + https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/base/webqtlDataset.py#L140-L169 + """ + fn_map = { + "ProbeSet": retrieve_probeset_trait_dataset_name, + "Publish": retrieve_publish_trait_dataset_name, + "Geno": retrieve_geno_trait_dataset_name, + "Temp": retrieve_temp_trait_dataset_name} + if trait_type == "Temp": + return retrieve_temp_trait_dataset_name(threshold, trait_name, conn) + return fn_map[trait_type](threshold, dataset_name, conn) + + +def retrieve_geno_riset_fields(name, conn): + """ + Retrieve the RISet, and RISetID values for various Geno trait types. + """ + query = ( + "SELECT InbredSet.Name, InbredSet.Id " + "FROM InbredSet, GenoFreeze " + "WHERE GenoFreeze.InbredSetId = InbredSet.Id " + "AND GenoFreeze.Name = %(name)s") + with conn.cursor() as cursor: + cursor.execute(query, {"name": name}) + return dict(zip(["riset", "risetid"], cursor.fetchone())) + return {} + +def retrieve_publish_riset_fields(name, conn): + """ + Retrieve the RISet, and RISetID values for various Publish trait types. + """ + query = ( + "SELECT InbredSet.Name, InbredSet.Id " + "FROM InbredSet, PublishFreeze " + "WHERE PublishFreeze.InbredSetId = InbredSet.Id " + "AND PublishFreeze.Name = %(name)s") + with conn.cursor() as cursor: + cursor.execute(query, {"name": name}) + return dict(zip(["riset", "risetid"], cursor.fetchone())) + return {} + +def retrieve_probeset_riset_fields(name, conn): + """ + Retrieve the RISet, and RISetID values for various ProbeSet trait types. + """ + query = ( + "SELECT InbredSet.Name, InbredSet.Id " + "FROM InbredSet, ProbeSetFreeze, ProbeFreeze " + "WHERE ProbeFreeze.InbredSetId = InbredSet.Id " + "AND ProbeFreeze.Id = ProbeSetFreeze.ProbeFreezeId " + "AND ProbeSetFreeze.Name = %(name)s") + with conn.cursor() as cursor: + cursor.execute(query, {"name": name}) + return dict(zip(["riset", "risetid"], cursor.fetchone())) + return {} + +def retrieve_temp_riset_fields(name, conn): + """ + Retrieve the RISet, and RISetID values for `Temp` trait types. + """ + query = ( + "SELECT InbredSet.Name, InbredSet.Id " + "FROM InbredSet, Temp " + "WHERE Temp.InbredSetId = InbredSet.Id " + "AND Temp.Name = %(name)s") + with conn.cursor() as cursor: + cursor.execute(query, {"name": name}) + return dict(zip(["riset", "risetid"], cursor.fetchone())) + return {} + +def retrieve_riset_fields(trait_type, trait_name, dataset_info, conn): + """ + Retrieve the RISet, and RISetID values for various trait types. + """ + riset_fns_map = { + "Geno": retrieve_geno_riset_fields, + "Publish": retrieve_publish_riset_fields, + "ProbeSet": retrieve_probeset_riset_fields + } + + if trait_type == "Temp": + riset_info = retrieve_temp_riset_fields(trait_name, conn) + else: + riset_info = riset_fns_map[trait_type](dataset_info["dataset_name"], conn) + + return { + **dataset_info, + **riset_info, + "riset": ( + "BXD" if riset_info.get("riset") == "BXD300" + else riset_info.get("riset", "")) + } + +def retrieve_temp_trait_dataset(): + """ + Retrieve the dataset that relates to `Temp` traits + """ + # pylint: disable=[C0330] + return { + "searchfield": ["name", "description"], + "disfield": ["name", "description"], + "type": "Temp", + "dataset_id": 1, + "fullname": "Temporary Storage", + "shortname": "Temp" + } + +def retrieve_geno_trait_dataset(): + """ + Retrieve the dataset that relates to `Geno` traits + """ + # pylint: disable=[C0330] + return { + "searchfield": ["name", "chr"], + "disfield": ["name", "chr", "mb", "source2", "sequence"], + "type": "Geno" + } + +def retrieve_publish_trait_dataset(): + """ + Retrieve the dataset that relates to `Publish` traits + """ + # pylint: disable=[C0330] + return { + "searchfield": [ + "name", "post_publication_description", "abstract", "title", + "authors"], + "disfield": [ + "name", "pubmed_id", "pre_publication_description", + "post_publication_description", "original_description", + "pre_publication_abbreviation", "post_publication_abbreviation", + "lab_code", "submitter", "owner", "authorized_users", + "authors", "title", "abstract", "journal", "volume", "pages", + "month", "year", "sequence", "units", "comments"], + "type": "Publish" + } + +def retrieve_probeset_trait_dataset(): + """ + Retrieve the dataset that relates to `ProbeSet` traits + """ + # pylint: disable=[C0330] + return { + "searchfield": [ + "name", "description", "probe_target_description", "symbol", + "alias", "genbankid", "unigeneid", "omim", "refseq_transcriptid", + "probe_set_specificity", "probe_set_blat_score"], + "disfield": [ + "name", "symbol", "description", "probe_target_description", "chr", + "mb", "alias", "geneid", "genbankid", "unigeneid", "omim", + "refseq_transcriptid", "blatseq", "targetseq", "chipid", "comments", + "strand_probe", "strand_gene", "probe_set_target_region", + "proteinid", "probe_set_specificity", "probe_set_blat_score", + "probe_set_blat_mb_start", "probe_set_blat_mb_end", + "probe_set_strand", "probe_set_note_by_rw", "flag"], + "type": "ProbeSet" + } + +def retrieve_trait_dataset(trait_type, trait, threshold, conn): + """ + Retrieve the dataset that relates to a specific trait. + """ + dataset_fns = { + "Temp": retrieve_temp_trait_dataset, + "Geno": retrieve_geno_trait_dataset, + "Publish": retrieve_publish_trait_dataset, + "ProbeSet": retrieve_probeset_trait_dataset + } + dataset_name_info = { + "dataset_id": None, + "dataset_name": trait["db"]["dataset_name"], + **retrieve_dataset_name( + trait_type, threshold, trait["trait_name"], + trait["db"]["dataset_name"], conn) + } + riset = retrieve_riset_fields( + trait_type, trait["trait_name"], dataset_name_info, conn) + return { + "display_name": dataset_name_info["dataset_name"], + **dataset_name_info, + **dataset_fns[trait_type](), + **riset + } diff --git a/gn3/db/traits.py b/gn3/db/traits.py index 61bc94d..6ea24be 100644 --- a/gn3/db/traits.py +++ b/gn3/db/traits.py @@ -1,5 +1,7 @@ """This class contains functions relating to trait data manipulation""" -from typing import Any, Dict, Union +from typing import Any, Dict, Union, Sequence +from gn3.function_helpers import compose +from gn3.db.datasets import retrieve_trait_dataset def get_trait_csv_sample_data(conn: Any, @@ -41,6 +43,7 @@ def update_sample_data(conn: Any, count: Union[int, str]): """Given the right parameters, update sample-data from the relevant table.""" + # pylint: disable=[R0913, R0914] STRAIN_ID_SQL: str = "UPDATE Strain SET Name = %s WHERE Id = %s" PUBLISH_DATA_SQL: str = ("UPDATE PublishData SET value = %s " "WHERE StrainId = %s AND Id = %s") @@ -76,159 +79,590 @@ def update_sample_data(conn: Any, return (updated_strains, updated_published_data, updated_se_data, updated_n_strains) - -def retrieve_trait_dataset_name( - trait_type: str, threshold: int, name: str, connection: Any): - """ - Retrieve the name of a trait given the trait's name - - This is extracted from the `webqtlDataset.retrieveName` function as is - implemented at - https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/base/webqtlDataset.py#L140-L169 - """ - columns = "Id, Name, FullName, ShortName{}".format( - ", DataScale" if trait_type == "ProbeSet" else "") - query = ( - "SELECT {columns} " - "FROM {trait_type}Freeze " - "WHERE " - "public > %(threshold)s " - "AND " - "(Name = %(name)s OR FullName = %(name)s OR ShortName = %(name)s)").format( - columns=columns, trait_type=trait_type) - with connection.cursor() as cursor: - cursor.execute(query, {"threshold": threshold, "name": name}) - return cursor.fetchone() - -PUBLISH_TRAIT_INFO_QUERY = ( - "SELECT " - "PublishXRef.Id, Publication.PubMed_ID, " - "Phenotype.Pre_publication_description, " - "Phenotype.Post_publication_description, " - "Phenotype.Original_description, " - "Phenotype.Pre_publication_abbreviation, " - "Phenotype.Post_publication_abbreviation, " - "Phenotype.Lab_code, Phenotype.Submitter, Phenotype.Owner, " - "Phenotype.Authorized_Users, CAST(Publication.Authors AS BINARY), " - "Publication.Title, Publication.Abstract, Publication.Journal, " - "Publication.Volume, Publication.Pages, Publication.Month, " - "Publication.Year, PublishXRef.Sequence, Phenotype.Units, " - "PublishXRef.comments " - "FROM " - "PublishXRef, Publication, Phenotype, PublishFreeze " - "WHERE " - "PublishXRef.Id = %(trait_name)s AND " - "Phenotype.Id = PublishXRef.PhenotypeId AND " - "Publication.Id = PublishXRef.PublicationId AND " - "PublishXRef.InbredSetId = PublishFreeze.InbredSetId AND " - "PublishFreeze.Id =%(trait_dataset_id)s") - - def retrieve_publish_trait_info(trait_data_source: Dict[str, Any], conn: Any): """Retrieve trait information for type `Publish` traits. https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/base/webqtlTrait.py#L399-L421""" + keys = ( + "Id", "PubMed_ID", "Pre_publication_description", + "Post_publication_description", "Original_description", + "Pre_publication_abbreviation", "Post_publication_abbreviation", + "Lab_code", "Submitter", "Owner", "Authorized_Users", "Authors", + "Title", "Abstract", "Journal", "Volume", "Pages", "Month", "Year", + "Sequence", "Units", "comments") + columns = ( + "PublishXRef.Id, Publication.PubMed_ID, " + "Phenotype.Pre_publication_description, " + "Phenotype.Post_publication_description, " + "Phenotype.Original_description, " + "Phenotype.Pre_publication_abbreviation, " + "Phenotype.Post_publication_abbreviation, " + "Phenotype.Lab_code, Phenotype.Submitter, Phenotype.Owner, " + "Phenotype.Authorized_Users, CAST(Publication.Authors AS BINARY), " + "Publication.Title, Publication.Abstract, Publication.Journal, " + "Publication.Volume, Publication.Pages, Publication.Month, " + "Publication.Year, PublishXRef.Sequence, Phenotype.Units, " + "PublishXRef.comments") + query = ( + "SELECT " + "{columns} " + "FROM " + "PublishXRef, Publication, Phenotype, PublishFreeze " + "WHERE " + "PublishXRef.Id = %(trait_name)s AND " + "Phenotype.Id = PublishXRef.PhenotypeId AND " + "Publication.Id = PublishXRef.PublicationId AND " + "PublishXRef.InbredSetId = PublishFreeze.InbredSetId AND " + "PublishFreeze.Id =%(trait_dataset_id)s").format(columns=columns) with conn.cursor() as cursor: cursor.execute( - PUBLISH_TRAIT_INFO_QUERY, + query, { k:v for k, v in trait_data_source.items() if k in ["trait_name", "trait_dataset_id"] }) - return cursor.fetchone() - -PROBESET_TRAIT_INFO_QUERY = ( - "SELECT " - "ProbeSet.name, ProbeSet.symbol, ProbeSet.description, " - "ProbeSet.probe_target_description, ProbeSet.chr, ProbeSet.mb, " - "ProbeSet.alias, ProbeSet.geneid, ProbeSet.genbankid, ProbeSet.unigeneid, " - "ProbeSet.omim, ProbeSet.refseq_transcriptid, ProbeSet.blatseq, " - "ProbeSet.targetseq, ProbeSet.chipid, ProbeSet.comments, " - "ProbeSet.strand_probe, ProbeSet.strand_gene, " - "ProbeSet.probe_set_target_region, ProbeSet.proteinid, " - "ProbeSet.probe_set_specificity, ProbeSet.probe_set_blat_score, " - "ProbeSet.probe_set_blat_mb_start, ProbeSet.probe_set_blat_mb_end, " - "ProbeSet.probe_set_strand, ProbeSet.probe_set_note_by_rw, " - "ProbeSet.flag " - "FROM " - "ProbeSet, ProbeSetFreeze, ProbeSetXRef " - "WHERE " - "ProbeSetXRef.ProbeSetFreezeId = ProbeSetFreeze.Id AND " - "ProbeSetXRef.ProbeSetId = ProbeSet.Id AND " - "ProbeSetFreeze.Name = %(trait_dataset_name)s AND " - "ProbeSet.Name = %(trait_name)s") + return dict(zip([k.lower() for k in keys], cursor.fetchone())) + +def set_confidential_field(trait_type, trait_info): + """Post processing function for 'Publish' trait types. + + It sets the value for the 'confidential' key.""" + if trait_type == "Publish": + return { + **trait_info, + "confidential": 1 if ( + trait_info.get("pre_publication_description", None) + and not trait_info.get("pubmed_id", None)) else 0} + return trait_info def retrieve_probeset_trait_info(trait_data_source: Dict[str, Any], conn: Any): """Retrieve trait information for type `ProbeSet` traits. https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/base/webqtlTrait.py#L424-L435""" + keys = ( + "name", "symbol", "description", "probe_target_description", "chr", + "mb", "alias", "geneid", "genbankid", "unigeneid", "omim", + "refseq_transcriptid", "blatseq", "targetseq", "chipid", "comments", + "strand_probe", "strand_gene", "probe_set_target_region", "proteinid", + "probe_set_specificity", "probe_set_blat_score", + "probe_set_blat_mb_start", "probe_set_blat_mb_end", "probe_set_strand", + "probe_set_note_by_rw", "flag") + query = ( + "SELECT " + "{columns} " + "FROM " + "ProbeSet, ProbeSetFreeze, ProbeSetXRef " + "WHERE " + "ProbeSetXRef.ProbeSetFreezeId = ProbeSetFreeze.Id AND " + "ProbeSetXRef.ProbeSetId = ProbeSet.Id AND " + "ProbeSetFreeze.Name = %(trait_dataset_name)s AND " + "ProbeSet.Name = %(trait_name)s").format( + columns=", ".join(["ProbeSet.{}".format(x) for x in keys])) with conn.cursor() as cursor: cursor.execute( - PROBESET_TRAIT_INFO_QUERY, + query, { k:v for k, v in trait_data_source.items() if k in ["trait_name", "trait_dataset_name"] }) - return cursor.fetchone() - -GENO_TRAIT_INFO_QUERY = ( - "SELECT " - "Geno.name, Geno.chr, Geno.mb, Geno.source2, Geno.sequence " - "FROM " - "Geno, GenoFreeze, GenoXRef " - "WHERE " - "GenoXRef.GenoFreezeId = GenoFreeze.Id AND GenoXRef.GenoId = Geno.Id AND " - "GenoFreeze.Name = %(trait_dataset_name)s AND Geno.Name = %(trait_name)s") + return dict(zip(keys, cursor.fetchone())) def retrieve_geno_trait_info(trait_data_source: Dict[str, Any], conn: Any): """Retrieve trait information for type `Geno` traits. https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/base/webqtlTrait.py#L438-L449""" + keys = ("name", "chr", "mb", "source2", "sequence") + query = ( + "SELECT " + "{columns} " + "FROM " + "Geno, GenoFreeze, GenoXRef " + "WHERE " + "GenoXRef.GenoFreezeId = GenoFreeze.Id AND GenoXRef.GenoId = Geno.Id AND " + "GenoFreeze.Name = %(trait_dataset_name)s AND " + "Geno.Name = %(trait_name)s").format( + columns=", ".join(["Geno.{}".format(x) for x in keys])) with conn.cursor() as cursor: cursor.execute( - GENO_TRAIT_INFO_QUERY, + query, { k:v for k, v in trait_data_source.items() if k in ["trait_name", "trait_dataset_name"] }) - return cursor.fetchone() - -TEMP_TRAIT_INFO_QUERY = ( - "SELECT name, description FROM Temp " - "WHERE Name = %(trait_name)s") + return dict(zip(keys, cursor.fetchone())) def retrieve_temp_trait_info(trait_data_source: Dict[str, Any], conn: Any): """Retrieve trait information for type `Temp` traits. https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/base/webqtlTrait.py#L450-452""" + keys = ("name", "description") + query = ( + "SELECT {columns} FROM Temp " + "WHERE Name = %(trait_name)s").format(columns=", ".join(keys)) with conn.cursor() as cursor: cursor.execute( - TEMP_TRAIT_INFO_QUERY, + query, { k:v for k, v in trait_data_source.items() if k in ["trait_name"] }) - return cursor.fetchone() + return dict(zip(keys, cursor.fetchone())) + +def set_haveinfo_field(trait_info): + """ + Common postprocessing function for all trait types. + + Sets the value for the 'haveinfo' field.""" + return {**trait_info, "haveinfo": 1 if trait_info else 0} + +def set_homologene_id_field_probeset(trait_info, conn): + """ + Postprocessing function for 'ProbeSet' traits. + + Sets the value for the 'homologene' key. + """ + query = ( + "SELECT HomologeneId FROM Homologene, Species, InbredSet" + " WHERE Homologene.GeneId = %(geneid)s AND InbredSet.Name = %(riset)s" + " AND InbredSet.SpeciesId = Species.Id AND" + " Species.TaxonomyId = Homologene.TaxonomyId") + with conn.cursor() as cursor: + cursor.execute( + query, + { + k:v for k, v in trait_info.items() + if k in ["geneid", "riset"] + }) + res = cursor.fetchone() + if res: + return {**trait_info, "homologeneid": res[0]} + return {**trait_info, "homologeneid": None} + +def set_homologene_id_field(trait_type, trait_info, conn): + """ + Common postprocessing function for all trait types. + + Sets the value for the 'homologene' key.""" + set_to_null = lambda ti: {**ti, "homologeneid": None} + functions_table = { + "Temp": set_to_null, + "Geno": set_to_null, + "Publish": set_to_null, + "ProbeSet": lambda ti: set_homologene_id_field_probeset(ti, conn) + } + return functions_table[trait_type](trait_info) + +def load_publish_qtl_info(trait_info, conn): + """ + Load extra QTL information for `Publish` traits + """ + query = ( + "SELECT PublishXRef.Locus, PublishXRef.LRS, PublishXRef.additive " + "FROM PublishXRef, PublishFreeze " + "WHERE PublishXRef.Id = %(trait_name)s " + "AND PublishXRef.InbredSetId = PublishFreeze.InbredSetId " + "AND PublishFreeze.Id = %(dataset_id)s") + with conn.cursor() as cursor: + cursor.execute( + query, + { + "trait_name": trait_info["trait_name"], + "dataset_id": trait_info["db"]["dataset_id"] + }) + return dict(zip(["locus", "lrs", "additive"], cursor.fetchone())) + return {"locus": "", "lrs": "", "additive": ""} + +def load_probeset_qtl_info(trait_info, conn): + """ + Load extra QTL information for `ProbeSet` traits + """ + query = ( + "SELECT ProbeSetXRef.Locus, ProbeSetXRef.LRS, ProbeSetXRef.pValue, " + "ProbeSetXRef.mean, ProbeSetXRef.additive " + "FROM ProbeSetXRef, ProbeSet " + "WHERE ProbeSetXRef.ProbeSetId = ProbeSet.Id " + " AND ProbeSet.Name = %(trait_name)s " + "AND ProbeSetXRef.ProbeSetFreezeId = %(dataset_id)s") + with conn.cursor() as cursor: + cursor.execute( + query, + { + "trait_name": trait_info["trait_name"], + "dataset_id": trait_info["db"]["dataset_id"] + }) + return dict(zip( + ["locus", "lrs", "pvalue", "mean", "additive"], cursor.fetchone())) + return {"locus": "", "lrs": "", "pvalue": "", "mean": "", "additive": ""} + +def load_qtl_info(qtl, trait_type, trait_info, conn): + """ + Load extra QTL information for traits + + DESCRIPTION: + Migrated from + https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/base/webqtlTrait.py#L500-L534 + + PARAMETERS: + qtl: boolean + trait_type: string + The type of the trait in consideration + trait_info: map/dictionary + A dictionary of the trait's key-value pairs + conn: + A database connection object + """ + if not qtl: + return trait_info + qtl_info_functions = { + "Publish": load_publish_qtl_info, + "ProbeSet": load_probeset_qtl_info + } + if trait_info["name"] not in qtl_info_functions.keys(): + return trait_info + + return qtl_info_functions[trait_type](trait_info, conn) + +def build_trait_name(trait_fullname): + """ + Initialises the trait's name, and other values from the search data provided + """ + def dataset_type(dset_name): + if dset_name.find('Temp') >= 0: + return "Temp" + if dset_name.find('Geno') >= 0: + return "Geno" + if dset_name.find('Publish') >= 0: + return "Publish" + return "ProbeSet" + + name_parts = trait_fullname.split("::") + assert len(name_parts) >= 2, "Name format error" + dataset_name = name_parts[0] + dataset_type = dataset_type(dataset_name) + return { + "db": { + "dataset_name": dataset_name, + "dataset_type": dataset_type}, + "trait_fullname": trait_fullname, + "trait_name": name_parts[1], + "cellid": name_parts[2] if len(name_parts) == 3 else "" + } + +def retrieve_probeset_sequence(trait, conn): + """ + Retrieve a 'ProbeSet' trait's sequence information + """ + query = ( + "SELECT ProbeSet.BlatSeq " + "FROM ProbeSet, ProbeSetFreeze, ProbeSetXRef " + "WHERE ProbeSet.Id=ProbeSetXRef.ProbeSetId " + "AND ProbeSetFreeze.Id = ProbeSetXRef.ProbeSetFreezeId " + "AND ProbeSet.Name = %(trait_name)s " + "AND ProbeSetFreeze.Name = %(dataset_name)s") + with conn.cursor() as cursor: + cursor.execute( + query, + { + "trait_name": trait["trait_name"], + "dataset_name": trait["db"]["dataset_name"] + }) + seq = cursor.fetchone() + return {**trait, "sequence": seq[0] if seq else ""} def retrieve_trait_info( - trait_type: str, trait_name: str, trait_dataset_id: int, - trait_dataset_name: str, conn: Any): + threshold: int, trait_full_name: str, conn: Any, + qtl=None): """Retrieves the trait information. https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/base/webqtlTrait.py#L397-L456 This function, or the dependent functions, might be incomplete as they are currently.""" + trait = build_trait_name(trait_full_name) + trait_dataset_type = trait["db"]["dataset_type"] trait_info_function_table = { "Publish": retrieve_publish_trait_info, "ProbeSet": retrieve_probeset_trait_info, "Geno": retrieve_geno_trait_info, "Temp": retrieve_temp_trait_info } - return trait_info_function_table[trait_type]( + + common_post_processing_fn = compose( + lambda ti: load_qtl_info(qtl, trait_dataset_type, ti, conn), + lambda ti: set_homologene_id_field(trait_dataset_type, ti, conn), + lambda ti: {"trait_type": trait_dataset_type, **ti}, + lambda ti: {**trait, **ti}) + + trait_post_processing_functions_table = { + "Publish": compose( + lambda ti: set_confidential_field(trait_dataset_type, ti), + common_post_processing_fn), + "ProbeSet": compose( + lambda ti: retrieve_probeset_sequence(ti, conn), + common_post_processing_fn), + "Geno": common_post_processing_fn, + "Temp": common_post_processing_fn + } + + retrieve_info = compose( + set_haveinfo_field, trait_info_function_table[trait_dataset_type]) + + trait_dataset = retrieve_trait_dataset( + trait_dataset_type, trait, threshold, conn) + trait_info = retrieve_info( { - "trait_name": trait_name, - "trait_dataset_id": trait_dataset_id, - "trait_dataset_name":trait_dataset_name + "trait_name": trait["trait_name"], + "trait_dataset_id": trait_dataset["dataset_id"], + "trait_dataset_name": trait_dataset["dataset_name"] }, conn) + if trait_info["haveinfo"]: + return { + **trait_post_processing_functions_table[trait_dataset_type](trait_info), + "db": {**trait["db"], **trait_dataset}, + "riset": trait_dataset["riset"] + } + return trait_info + +def retrieve_temp_trait_data(trait_info: dict, conn: Any): + """ + Retrieve trait data for `Temp` traits. + """ + query = ( + "SELECT " + "Strain.Name, TempData.value, TempData.SE, TempData.NStrain, " + "TempData.Id " + "FROM TempData, Temp, Strain " + "WHERE TempData.StrainId = Strain.Id " + "AND TempData.Id = Temp.DataId " + "AND Temp.name = %(trait_name)s " + "ORDER BY Strain.Name") + with conn.cursor() as cursor: + cursor.execute( + query, + {"trait_name": trait_info["trait_name"]}) + return [dict(zip( + ["strain_name", "value", "se_error", "nstrain", "id"], row)) + for row in cursor.fetchall()] + return [] + +def retrieve_species_id(riset, conn: Any): + """ + Retrieve a species id given the RISet value + """ + with conn.cursor as cursor: + cursor.execute( + "SELECT SpeciesId from InbredSet WHERE Name = %(riset)s", + {"riset": riset}) + return cursor.fetchone()[0] + return None + +def retrieve_geno_trait_data(trait_info: Dict, conn: Any): + """ + Retrieve trait data for `Geno` traits. + """ + query = ( + "SELECT Strain.Name, GenoData.value, GenoSE.error, GenoData.Id " + "FROM (GenoData, GenoFreeze, Strain, Geno, GenoXRef) " + "LEFT JOIN GenoSE ON " + "(GenoSE.DataId = GenoData.Id AND GenoSE.StrainId = GenoData.StrainId) " + "WHERE Geno.SpeciesId = %(species_id)s " + "AND Geno.Name = %(trait_name)s AND GenoXRef.GenoId = Geno.Id " + "AND GenoXRef.GenoFreezeId = GenoFreeze.Id " + "AND GenoFreeze.Name = %(dataset_name)s " + "AND GenoXRef.DataId = GenoData.Id " + "AND GenoData.StrainId = Strain.Id " + "ORDER BY Strain.Name") + with conn.cursor() as cursor: + cursor.execute( + query, + {"trait_name": trait_info["trait_name"], + "dataset_name": trait_info["db"]["dataset_name"], + "species_id": retrieve_species_id( + trait_info["db"]["riset"], conn)}) + return [dict(zip( + ["strain_name", "value", "se_error", "id"], row)) + for row in cursor.fetchall()] + return [] + +def retrieve_publish_trait_data(trait_info: Dict, conn: Any): + """ + Retrieve trait data for `Publish` traits. + """ + query = ( + "SELECT " + "Strain.Name, PublishData.value, PublishSE.error, NStrain.count, " + "PublishData.Id " + "FROM (PublishData, Strain, PublishXRef, PublishFreeze) " + "LEFT JOIN PublishSE ON " + "(PublishSE.DataId = PublishData.Id " + "AND PublishSE.StrainId = PublishData.StrainId) " + "LEFT JOIN NStrain ON " + "(NStrain.DataId = PublishData.Id " + "AND NStrain.StrainId = PublishData.StrainId) " + "WHERE PublishXRef.InbredSetId = PublishFreeze.InbredSetId " + "AND PublishData.Id = PublishXRef.DataId " + "AND PublishXRef.Id = %(trait_name)s " + "AND PublishFreeze.Id = %(dataset_id)s " + "AND PublishData.StrainId = Strain.Id " + "ORDER BY Strain.Name") + with conn.cursor() as cursor: + cursor.execute( + query, + {"trait_name": trait_info["trait_name"], + "dataset_id": trait_info["db"]["dataset_id"]}) + return [dict(zip( + ["strain_name", "value", "se_error", "nstrain", "id"], row)) + for row in cursor.fetchall()] + return [] + +def retrieve_cellid_trait_data(trait_info: Dict, conn: Any): + """ + Retrieve trait data for `Probe Data` types. + """ + query = ( + "SELECT " + "Strain.Name, ProbeData.value, ProbeSE.error, ProbeData.Id " + "FROM (ProbeData, ProbeFreeze, ProbeSetFreeze, ProbeXRef, Strain," + " Probe, ProbeSet) " + "LEFT JOIN ProbeSE ON " + "(ProbeSE.DataId = ProbeData.Id " + " AND ProbeSE.StrainId = ProbeData.StrainId) " + "WHERE Probe.Name = %(cellid)s " + "AND ProbeSet.Name = %(trait_name)s " + "AND Probe.ProbeSetId = ProbeSet.Id " + "AND ProbeXRef.ProbeId = Probe.Id " + "AND ProbeXRef.ProbeFreezeId = ProbeFreeze.Id " + "AND ProbeSetFreeze.ProbeFreezeId = ProbeFreeze.Id " + "AND ProbeSetFreeze.Name = %(dataset_name)s " + "AND ProbeXRef.DataId = ProbeData.Id " + "AND ProbeData.StrainId = Strain.Id " + "ORDER BY Strain.Name") + with conn.cursor() as cursor: + cursor.execute( + query, + {"cellid": trait_info["cellid"], + "trait_name": trait_info["trait_name"], + "dataset_id": trait_info["db"]["dataset_id"]}) + return [dict(zip( + ["strain_name", "value", "se_error", "id"], row)) + for row in cursor.fetchall()] + return [] + +def retrieve_probeset_trait_data(trait_info: Dict, conn: Any): + """ + Retrieve trait data for `ProbeSet` traits. + """ + query = ( + "SELECT Strain.Name, ProbeSetData.value, ProbeSetSE.error, " + "ProbeSetData.Id " + "FROM (ProbeSetData, ProbeSetFreeze, Strain, ProbeSet, ProbeSetXRef) " + "LEFT JOIN ProbeSetSE ON " + "(ProbeSetSE.DataId = ProbeSetData.Id " + "AND ProbeSetSE.StrainId = ProbeSetData.StrainId) " + "WHERE ProbeSet.Name = %(trait_name)s " + "AND ProbeSetXRef.ProbeSetId = ProbeSet.Id " + "AND ProbeSetXRef.ProbeSetFreezeId = ProbeSetFreeze.Id " + "AND ProbeSetFreeze.Name = %(dataset_name)s " + "AND ProbeSetXRef.DataId = ProbeSetData.Id " + "AND ProbeSetData.StrainId = Strain.Id " + "ORDER BY Strain.Name") + + with conn.cursor() as cursor: + cursor.execute( + query, + {"trait_name": trait_info["trait_name"], + "dataset_name": trait_info["db"]["dataset_name"]}) + return [dict(zip( + ["strain_name", "value", "se_error", "id"], row)) + for row in cursor.fetchall()] + return [] + +def with_strainlist_data_setup(strainlist: Sequence[str]): + """ + Build function that computes the trait data from provided list of strains. + + PARAMETERS + strainlist: (list) + A list of strain names + + RETURNS: + Returns a function that given some data from the database, computes the + strain's value, variance and ndata values, only if the strain is present + in the provided `strainlist` variable. + """ + def setup_fn(tdata): + if tdata["strain_name"] in strainlist: + val = tdata["value"] + if val is not None: + return { + "strain_name": tdata["strain_name"], + "value": val, + "variance": tdata["se_error"], + "ndata": tdata.get("nstrain", None) + } + return None + return setup_fn + +def without_strainlist_data_setup(): + """ + Build function that computes the trait data. + + RETURNS: + Returns a function that given some data from the database, computes the + strain's value, variance and ndata values. + """ + def setup_fn(tdata): + val = tdata["value"] + if val is not None: + return { + "strain_name": tdata["strain_name"], + "value": val, + "variance": tdata["se_error"], + "ndata": tdata.get("nstrain", None) + } + return None + return setup_fn + +def retrieve_trait_data(trait: dict, conn: Any, strainlist: Sequence[str] = tuple()): + """ + Retrieve trait data + + DESCRIPTION + Retrieve trait data as is done in + https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/base/webqtlTrait.py#L258-L386 + """ + # I do not like this section, but it retains the flow in the old codebase + if trait["db"]["dataset_type"] == "Temp": + results = retrieve_temp_trait_data(trait, conn) + elif trait["db"]["dataset_type"] == "Publish": + results = retrieve_publish_trait_data(trait, conn) + elif trait["cellid"]: + results = retrieve_cellid_trait_data(trait, conn) + elif trait["db"]["dataset_type"] == "ProbeSet": + results = retrieve_probeset_trait_data(trait, conn) + else: + results = retrieve_geno_trait_data(trait, conn) + + if results: + # do something with mysqlid + mysqlid = results[0]["id"] + if strainlist: + data = [ + item for item in + map(with_strainlist_data_setup(strainlist), results) + if item is not None] + else: + data = [ + item for item in + map(without_strainlist_data_setup(), results) + if item is not None] + + return { + "mysqlid": mysqlid, + "data": dict(map( + lambda x: ( + x["strain_name"], + {k:v for k, v in x.items() if x != "strain_name"}), + data))} + return {} diff --git a/gn3/function_helpers.py b/gn3/function_helpers.py new file mode 100644 index 0000000..397b2da --- /dev/null +++ b/gn3/function_helpers.py @@ -0,0 +1,36 @@ +""" +This module will contain helper functions that should assist in maintaining a +mostly functional way of programming. + +It will also contain miscellaneous functions that can be used globally, and thus +do not fit well in any other module. + +FUNCTIONS: +compose: This function is used to compose multiple functions into a single + function. It passes the results of calling one function to the other until + all the functions to be composed are called. +""" +from functools import reduce + +def compose(*functions): + """Compose multiple functions into a single function. + + The utility in this function is not specific to this module, and as such, + this function can, and probably should, be moved to a more global module. + + DESCRIPTION: + Given `cfn = compose(f_1, f_2, ... f_(n-1), f_n )`, calling + `cfn(arg_1, arg_2, ..., arg_m)` should call `f_n` with the arguments passed + to `cfn` and the results of that should be passed as arguments to `f_(n-1)` + and so on until `f_1` is called with the results of the cumulative calls and + that is the result of the entire chain of calls. + + PARAMETERS: + functions: a variable argument list of function. + """ + def composed_function(*args, **kwargs): + return reduce( + lambda res, fn: fn(res), + reversed(functions[:-1]), + functions[-1](*args, **kwargs)) + return composed_function |