From 31ac939f58bf7b6d353ced995ca395376203b25f Mon Sep 17 00:00:00 2001 From: Alexander Kabui Date: Mon, 12 Apr 2021 09:54:12 +0300 Subject: Integrate correlation API - add new api for gn2-gn3 sample r integration - delete map for sample list to values - add db util file - add python msql-client dependency - add db for fetching lit correlation results - add unittests for db utils - add tests for db_utils - modify api for fetching lit correlation results - refactor Mock Database Connector and unittests - add sql url parser - add SQL URI env variable - refactor code for db utils - modify return data for lit correlation - refactor tissue correlation endpoint - replace db_instance with conn--- gn3/api/correlation.py | 39 +++++++++++---- gn3/computations/correlations.py | 104 +++++++++++++++++++++++++++++---------- gn3/db_utils.py | 24 +++++++++ gn3/settings.py | 2 +- 4 files changed, 132 insertions(+), 37 deletions(-) create mode 100644 gn3/db_utils.py (limited to 'gn3') diff --git a/gn3/api/correlation.py b/gn3/api/correlation.py index 53ea6a7..2339088 100644 --- a/gn3/api/correlation.py +++ b/gn3/api/correlation.py @@ -1,6 +1,4 @@ """Endpoints for running correlations""" -from unittest import mock - from flask import jsonify from flask import Blueprint from flask import request @@ -8,11 +6,31 @@ from flask import request from gn3.computations.correlations import compute_all_sample_correlation from gn3.computations.correlations import compute_all_lit_correlation from gn3.computations.correlations import compute_all_tissue_correlation - +from gn3.computations.correlations import map_shared_keys_to_values +from gn3.db_utils import database_connector correlation = Blueprint("correlation", __name__) +@correlation.route("/sample_x/", methods=["POST"]) +def compute_sample_integration(corr_method="pearson"): + """temporary api to help integrate genenetwork2 to genenetwork3 """ + + correlation_input = request.get_json() + + target_samplelist = correlation_input.get("target_samplelist") + target_data_values = correlation_input.get("target_dataset") + this_trait_data = correlation_input.get("trait_data") + + results = map_shared_keys_to_values(target_samplelist, target_data_values) + + correlation_results = compute_all_sample_correlation(corr_method=corr_method, + this_trait=this_trait_data, + target_dataset=results) + + return jsonify(correlation_results) + + @correlation.route("/sample_r/", methods=["POST"]) def compute_sample_r(corr_method="pearson"): """correlation endpoint for computing sample r correlations\ @@ -22,11 +40,11 @@ def compute_sample_r(corr_method="pearson"): # xtodo move code below to compute_all_sampl correlation this_trait_data = correlation_input.get("this_trait") - target_datasets = correlation_input.get("target_dataset") + target_dataset_data = correlation_input.get("target_dataset") correlation_results = compute_all_sample_correlation(corr_method=corr_method, this_trait=this_trait_data, - target_dataset=target_datasets) + target_dataset=target_dataset_data) return jsonify({ "corr_results": correlation_results @@ -39,13 +57,16 @@ def compute_lit_corr(species=None, gene_id=None): are fetched from the database this is the only case where the db\ might be needed for actual computing of the correlation results""" - database_instance = mock.Mock() + conn, _cursor_object = database_connector() target_traits_gene_ids = request.get_json() + target_trait_gene_list = list(target_traits_gene_ids.items()) lit_corr_results = compute_all_lit_correlation( - database_instance=database_instance, trait_lists=target_traits_gene_ids, + conn=conn, trait_lists=target_trait_gene_list, species=species, gene_id=gene_id) + conn.close() + return jsonify(lit_corr_results) @@ -54,10 +75,10 @@ def compute_tissue_corr(corr_method="pearson"): """api endpoint fr doing tissue correlation""" tissue_input_data = request.get_json() primary_tissue_dict = tissue_input_data["primary_tissue"] - target_tissues_dict_list = tissue_input_data["target_tissues"] + target_tissues_dict = tissue_input_data["target_tissues_dict"] results = compute_all_tissue_correlation(primary_tissue_dict=primary_tissue_dict, - target_tissues_dict_list=target_tissues_dict_list, + target_tissues_data=target_tissues_dict, corr_method=corr_method) return jsonify(results) diff --git a/gn3/computations/correlations.py b/gn3/computations/correlations.py index dc2f8d3..26b7294 100644 --- a/gn3/computations/correlations.py +++ b/gn3/computations/correlations.py @@ -12,10 +12,30 @@ def compute_sum(rhs: int, lhs: int) -> int: return rhs + lhs +def map_shared_keys_to_values(target_sample_keys: List, target_sample_vals: dict)-> List: + """Function to construct target dataset data items given commoned shared\ + keys and trait samplelist values for example given keys >>>>>>>>>>\ + ["BXD1", "BXD2", "BXD5", "BXD6", "BXD8", "BXD9"] and value object as\ + "HCMA:_AT": [4.1, 5.6, 3.2, 1.1, 4.4, 2.2],TXD_AT": [6.2, 5.7, 3.6, 1.5, 4.2, 2.3]}\ + return results should be a list of dicts mapping the shared keys to the trait values""" + target_dataset_data = [] + + for trait_id, sample_values in target_sample_vals.items(): + target_trait_dict = dict(zip(target_sample_keys, sample_values)) + + target_trait = { + "trait_id": trait_id, + "trait_sample_data": target_trait_dict + } + + target_dataset_data.append(target_trait) + + return target_dataset_data + + def normalize_values(a_values: List, b_values: List) -> Tuple[List[float], List[float], int]: """Trim two lists of values to contain only the values they both share - Given two lists of sample values, trim each list so that it contains only the samples that contain a value in both lists. Also returns the number of such samples. @@ -175,7 +195,7 @@ def tissue_correlation_for_trait_list( """ - # ax :todo assertion that lenggth one one target tissue ==primary_tissue + # ax :todo assertion that length one one target tissue ==primary_tissue (tissue_corr_coeffient, p_value) = compute_corr_p_value(primary_values=primary_tissue_vals, @@ -192,11 +212,11 @@ def tissue_correlation_for_trait_list( def fetch_lit_correlation_data( - database, + conn, input_mouse_gene_id: Optional[str], gene_id: str, mouse_gene_id: Optional[str] = None) -> Tuple[str, float]: - """given input trait mouse gene id and mouse gene id fetch the lit\ + """Given input trait mouse gene id and mouse gene id fetch the lit\ corr_data""" if mouse_gene_id is not None and ";" not in mouse_gene_id: query = """ @@ -208,15 +228,19 @@ def fetch_lit_correlation_data( query_values = (str(mouse_gene_id), str(input_mouse_gene_id)) - results = database.execute(query_formatter(query, - *query_values)).fetchone() + cursor = conn.cursor() + + cursor.execute(query_formatter(query, + *query_values)) + results = cursor.fetchone() lit_corr_results = None if results is not None: lit_corr_results = results else: - lit_corr_results = database.execute( - query_formatter(query, - *tuple(reversed(query_values)))).fetchone() + cursor = conn.cursor() + cursor.execute(query_formatter(query, + *tuple(reversed(query_values)))) + lit_corr_results = cursor.fetchone() lit_results = (gene_id, lit_corr_results.val)\ if lit_corr_results else (gene_id, 0) return lit_results @@ -225,7 +249,7 @@ def fetch_lit_correlation_data( def lit_correlation_for_trait_list( - database, + conn, target_trait_lists: List, species: Optional[str] = None, trait_gene_id: Optional[str] = None) -> List: @@ -233,43 +257,45 @@ def lit_correlation_for_trait_list( output is float for lit corr results """ fetched_lit_corr_results = [] - this_trait_mouse_gene_id = map_to_mouse_gene_id(database=database, + this_trait_mouse_gene_id = map_to_mouse_gene_id(conn=conn, species=species, gene_id=trait_gene_id) - for trait in target_trait_lists: - target_trait_gene_id = trait.get("gene_id") + for (trait_name, target_trait_gene_id) in target_trait_lists: + corr_results = {} if target_trait_gene_id: target_mouse_gene_id = map_to_mouse_gene_id( - database=database, + conn=conn, species=species, gene_id=target_trait_gene_id) fetched_corr_data = fetch_lit_correlation_data( - database=database, + conn=conn, input_mouse_gene_id=this_trait_mouse_gene_id, gene_id=target_trait_gene_id, mouse_gene_id=target_mouse_gene_id) dict_results = dict(zip(("gene_id", "lit_corr"), fetched_corr_data)) - fetched_lit_corr_results.append(dict_results) + corr_results[trait_name] = dict_results + fetched_lit_corr_results.append(corr_results) return fetched_lit_corr_results def query_formatter(query_string: str, *query_values): - """formatter query string given the unformatted query string\ + """Formatter query string given the unformatted query string\ and the respectibe values.Assumes number of placeholders is equal to the number of query values """ + # xtodo escape sql queries results = query_string % (query_values) return results -def map_to_mouse_gene_id(database, species: Optional[str], +def map_to_mouse_gene_id(conn, species: Optional[str], gene_id: Optional[str]) -> Optional[str]: - """given a species which is not mouse map the gene_id\ + """Given a species which is not mouse map the gene_id\ to respective mouse gene id""" # AK:xtodo move the code for checking nullity out of thing functions bug # while method for string @@ -278,28 +304,29 @@ def map_to_mouse_gene_id(database, species: Optional[str], if species == "mouse": return gene_id + cursor = conn.cursor() query = """SELECT mouse FROM GeneIDXRef WHERE '%s' = '%s'""" query_values = (species, gene_id) - - results = database.execute(query_formatter(query, - *query_values)).fetchone() + cursor.execute(query_formatter(query, + *query_values)) + results = cursor.fetchone() mouse_gene_id = results.mouse if results is not None else None return mouse_gene_id -def compute_all_lit_correlation(database_instance, trait_lists: List, +def compute_all_lit_correlation(conn, trait_lists: List, species: str, gene_id): """Function that acts as an abstraction for lit_correlation_for_trait_list""" # xtodo to be refactored lit_results = lit_correlation_for_trait_list( - database=database_instance, + conn=conn, target_trait_lists=trait_lists, species=species, trait_gene_id=gene_id) @@ -308,18 +335,22 @@ def compute_all_lit_correlation(database_instance, trait_lists: List, def compute_all_tissue_correlation(primary_tissue_dict: dict, - target_tissues_dict_list: List, + target_tissues_data: dict, corr_method: str): """Function acts as an abstraction for tissue_correlation_for_trait_list\ - required input are target tissue object and primary tissue trait + required input are target tissue object and primary tissue trait\ + target tissues data contains the trait_symbol_dict and symbol_tissue_vals """ tissues_results = {} primary_tissue_vals = primary_tissue_dict["tissue_values"] + traits_symbol_dict = target_tissues_data["trait_symbol_dict"] + symbol_tissue_vals_dict = target_tissues_data["symbol_tissue_vals_dict"] - target_tissues_list = target_tissues_dict_list + target_tissues_list = process_trait_symbol_dict( + traits_symbol_dict, symbol_tissue_vals_dict) for target_tissue_obj in target_tissues_list: trait_id = target_tissue_obj.get("trait_id") @@ -334,3 +365,22 @@ def compute_all_tissue_correlation(primary_tissue_dict: dict, tissues_results[trait_id] = tissue_result return tissues_results + + +def process_trait_symbol_dict(trait_symbol_dict, symbol_tissue_vals_dict) -> List: + """Method for processing trait symbol\ + dict given the symbol tissue values """ + traits_tissue_vals = [] + + for (trait, symbol) in trait_symbol_dict.items(): + if symbol is not None: + target_symbol = symbol.lower() + if target_symbol in symbol_tissue_vals_dict: + trait_tissue_val = symbol_tissue_vals_dict[target_symbol] + target_tissue_dict = {"trait_id": trait, + "symbol": target_symbol, + "tissue_values": trait_tissue_val} + + traits_tissue_vals.append(target_tissue_dict) + + return traits_tissue_vals diff --git a/gn3/db_utils.py b/gn3/db_utils.py new file mode 100644 index 0000000..34c5bf0 --- /dev/null +++ b/gn3/db_utils.py @@ -0,0 +1,24 @@ +"""module contains all db related stuff""" +from typing import Tuple +from urllib.parse import urlparse +import MySQLdb as mdb # type: ignore +from gn3.settings import SQL_URI + + +def parse_db_url() -> Tuple: + """function to parse SQL_URI env variable note:there\ + is a default value for SQL_URI so a tuple result is\ + always expected""" + parsed_db = urlparse(SQL_URI) + return (parsed_db.hostname, parsed_db.username, + parsed_db.password, parsed_db.path[1:]) + + +def database_connector()->Tuple: + """function to create db connector""" + host, user, passwd, db_name = parse_db_url() + conn = mdb.connect(host, user, passwd, db_name) + cursor = conn.cursor() + + return (conn, cursor) + \ No newline at end of file diff --git a/gn3/settings.py b/gn3/settings.py index 2836581..e77a977 100644 --- a/gn3/settings.py +++ b/gn3/settings.py @@ -12,6 +12,6 @@ REDIS_JOB_QUEUE = "GN3::job-queue" TMPDIR = os.environ.get("TMPDIR", tempfile.gettempdir()) # SQL confs -SQLALCHEMY_DATABASE_URI = "mysql://kabui:1234@localhost/test" +SQL_URI = os.environ.get("SQL_URI", "mysql://kabui:1234@localhost/db_webqtl") SECRET_KEY = "password" SQLALCHEMY_TRACK_MODIFICATIONS = False -- cgit v1.2.3