create sql script

dev/create_sc_expression_sql.py

@@ -0,0 +1,103 @@
+#!/usr/bin/env python
+"""
+Create a SQL script to insert single_cell_expression data into the database
+"""
+
+import json
+import pandas as pd
+
+# STUDY_ID = msk_spectrum_tme_2022
+cancer_study_id = 40
+# THIS DOESNT MATTER, JUST A HIGH NUMBER?
+genetic_profile_id = 10000
+
+
+create_table_statement = """
+DROP TABLE IF EXISTS single_cell_expression;
+CREATE TABLE IF NOT EXISTS single_cell_expression (
+    GENETIC_PROFILE_ID int NOT NULL,
+    SAMPLE_ID int NOT NULL,
+    TISSUE varchar(255) NOT NULL,
+    CELL_TYPE varchar(255) NOT NULL,
+    ENTREZ_GENE_ID int NOT NULL,
+    EXPRESSION_VALUE float,
+    FOREIGN KEY(GENETIC_PROFILE_ID) REFERENCES genetic_profile(GENETIC_PROFILE_ID),
+    FOREIGN KEY(SAMPLE_ID) REFERENCES sample(INTERNAL_ID),
+    FOREIGN KEY(ENTREZ_GENE_ID) REFERENCES gene(ENTREZ_GENE_ID)
+);
+"""
+
+add_genetic_profile = f"""
+DELETE FROM genetic_profile WHERE STABLE_ID = "single_cell_expression";
+INSERT INTO genetic_profile (
+    GENETIC_PROFILE_ID, STABLE_ID, CANCER_STUDY_ID, GENETIC_ALTERATION_TYPE,
+    DATATYPE, NAME, DESCRIPTION, SHOW_PROFILE_IN_ANALYSIS_TAB
+) VALUES (
+     {genetic_profile_id}, "SINGLE_CELL_EXPRESSION", {cancer_study_id}, "single_cell_expression",
+    "single_cell_expression", "Single Cell Expression", "Single Cell Expression", 1
+);
+"""
+
+insert_data_start = """
+INSERT INTO single_cell_expression (
+    GENETIC_PROFILE_ID, SAMPLE_ID, TISSUE, CELL_TYPE, ENTREZ_GENE_ID, EXPRESSION_VALUE
+) VALUES (
+"""
+
+value_split = "\n), (\n"
+
+def create_sample_map():
+    """SQL
+    SELECT sample.INTERNAL_ID, sample.STABLE_ID FROM sample 
+    INNER JOIN patient ON sample.PATIENT_ID = patient.INTERNAL_ID 
+    INNER JOIN cancer_study ON patient.CANCER_STUDY_ID = cancer_study.CANCER_STUDY_ID 
+    WHERE cancer_study.CANCER_STUDY_IDENTIFIER = "msk_spectrum_tme_2022";
+    """
+    data = pd.read_csv("sample_map.tsv", skiprows=1, sep="\t")
+    data = data.dropna(how="any")
+    return dict(zip(data["INTERNAL_ID"], data["STABLE_ID"]))
+
+
+def create_gene_map():
+    """SQL:
+    select ENTREZ_GENE_ID, HUGO_GENE_SYMBOL from gene;
+    """
+    data = pd.read_csv("gene_map.tsv", skiprows=1, sep="\t")
+    data = data.dropna(how="any")
+    return dict(zip(data["ENTREZ_GENE_ID"], data["HUGO_GENE_SYMBOL"]))
+
+
+def create_data_sql() -> str:
+    sample_map = create_sample_map()
+    gene_map = create_gene_map()
+    with open("sc-expression-msk-spectrum.json") as f:
+        data = json.loads(f.read())
+
+    sql = """"""
+    first = True
+    for sample_id in data.keys():
+        mapped_sample_id = sample_map.get(sample_id)
+        if mapped_sample_id is None:
+            continue
+        for tissue in data[sample_id].keys():
+            for cell_type in data[sample_id][tissue].keys():
+                for gene, value in data[sample_id][tissue][cell_type].items():
+                    mapped_gene_id = gene_map.get(gene)
+                    if mapped_gene_id is None:
+                        continue
+                    if not first:
+                        sql += "), ("
+                    sql += f"{genetic_profile_id}, {mapped_sample_id}, \"{tissue}\", \"{cell_type}\", {mapped_gene_id}, {value} "
+                    first = False
+    sql += ");"
+    print(sql)
+    return sql
+
+def create_sql_file():
+    with open("single_cell_expression.sql", "w+") as f:
+        f.write(create_table_statement + "\n")
+        f.write(add_genetic_profile + "\n")
+        f.write(insert_data_start)
+        f.write(create_data_sql())
+
+create_sql_file()