have the basics of a downloader and the sql to support it.

AACT_downloader/ClinicalTrialHistory/docker-entrypoint-initdb.d/020_HttpSchema.sql

@@ -22,20 +22,13 @@ As not every request will have an xml doc, split them.
 
 CREATE TABLE IF NOT EXISTS http.responses (
     id SERIAL PRIMARY KEY, 
-    nct VARCHAR(15),
-    version SMALLINT,
+    nct_id VARCHAR(15),
+    version_a SMALLINT,
+	version_b SMALLINT,
     url VARCHAR(255),
     response_code SMALLINT,
-    response_date DATE
+    response_date TIMESTAMP WITH TIME ZONE,
+	html TEXT
     );
 
-CREATE TABLE IF NOT EXISTS http.xml_documents (
-    id SERIAL PRIMARY KEY,
-    xml XML,
-    CONSTRAINT http_response 
-        FOREIGN KEY (id) 
-        REFERENCES http.responses (id)
-        ON DELETE CASCADE --remove xml if the request is deleted
-);
-
 

downloader/db_connection.py

@@ -4,8 +4,8 @@ conn = psyco.connect(dbname="aact_db", user="root", host="localhost", password="
 
 curse = conn.cursor()
 
-curse.execute("SELECT * FROM ctgov.studies LIMIT 2;")
+curse.execute("select * FROM http.responses")
 print(curse.fetchall())
 
 curse.close()
-conn.close()
+conn.close()

downloader/downloader.py

@@ -0,0 +1,96 @@
+import requests
+import psycopg2 as psyco
+from datetime import datetime
+from bs4 import BeautifulSoup
+
+
+def get_highest_version_number(response):
+    #navigate to a specific part of the returned html and extract the highest posted version.
+    soup = BeautifulSoup(response.text, features="lxml")
+    version_value = soup.findChildren("fieldset")[0].table.tbody.findChildren("tr")[-1].td.text
+    return int(version_value)
+
+def make_request(nct_id,version1,version2):
+    #create url
+    baseurl = "https://clinicaltrials.gov/ct2/history/{}?A={}&B={}&C=Side-by-Side"
+    url = baseurl.format(nct_id,version1,version2)
+
+    #make request
+    response = requests.get(url)
+
+    #return the response
+    return response
+
+def upload_response(db_conn, nct_id, version_a, version_b, response):
+
+    timestamp = datetime.strptime(r.headers['date'], "%a, %d %b %Y %H:%M:%S %Z")
+
+    #this uploads the response values.
+    db_conn.execute("""
+    INSERT INTO http.responses
+    (nct_id,version_a,version_b,url,response_code,response_date, html) 
+    VALUES  (%s,%s,%s,%s,%s,%s,%s)
+    RETURNING id
+    ;
+    """
+    ,(nct_id
+        ,version_a
+        ,version_b
+        ,response.url
+        ,response.status_code
+        ,datetime.isoformat(timestamp)
+        ,response.text)) #FIX: there is an issue with storing the tags for some reason.
+
+    pk_response = cursor.fetchall()
+
+    return pk_response
+
+
+def download_trial_records(nct_id, db_conn):
+    #download first 2 records
+    r = make_request(nct_id,1,2)
+    upload_response(db_conn, nct_id,1,2,r)
+
+    #extract last version
+    v = get_highest_version_number(r)
+
+    #download the remaining versions
+    if v == 2:
+        return None
+    elif v % 2 == 0:
+        for version_a, version_b in step_generator(v):
+            r = make_request(nct_id, version_a, version_b)
+            upload_response(db_conn,nct_id,version_a, version_b, r)
+
+    elif v % 2 == 1:
+        for version_a, version_b in step_generator(v):
+            r = make_request(nct_id, version_a, version_b)
+            upload_response(db_conn,nct_id,version_a, version_b, r)
+
+        r = make_request(nct_id, 1, v)
+        upload_response(db_conn,nct_id,1, v, r)
+
+def step_generator(max_version):
+    old=3
+    for i in range(4,max_version,2):
+        yield (old,i)
+        old = i + 1
+
+if __name__ == "__main__":
+
+    #make request
+    nct_id = "NCT00658567"
+    r = make_request(nct_id,1,36)
+    print(r.url)
+    print(r.status_code)
+
+    v = get_highest_version_number(r)
+    print("highest version", v )
+
+    #db connection
+    with psyco.connect(dbname="aact_db", user="root", host="localhost", password="root") as con:
+        with con.cursor() as cursor:
+            #insert response
+
+            download_trial_records(nct_id,cursor)
+            print("download complete")