Mostly working data extraction, removed dependency on text processing (will remove in future commit)

Parser/extraction-lib.py

@@ -1,79 +1,227 @@
-from tokenize import String
+from collections import namedtuple
+from copy import copy
+from datetime import datetime
+from ensurepip import version
 from bs4 import BeautifulSoup
 import abc
 import textprocessing as tp #cuz tp is important
 #requires Python 3.10
 
-def extract_data_from_tr(tr) -> tuple[String, String]:
-    """
-    Takes an html data row of interest, extracts the record_name from the first <td>, and the data from the second <td>.
 
-    For the data, it will split between old and new data, making copies of each and returnign them.
+###CLASSES AND CONSTRUCTORS
 
-    Uses functionality from ./textprocessing.py (separated because it is important to test that functionality)
+TagDatePair = namedtuple("TagDatePair", ["tag","date"])
-    to get extract data from tags.
+TagTextPair = namedtuple("TagTextPair", ["tag","text"])
-
-    """
-    #get list of cells
-    #for cell in cells
-    #if class_=="rowLabel", extract text
-    #else parse out new and old text
-    #return triple: row_lable, old, new
-    pass
 
 #superclasses
-class VersionData{abc.ABC}:
+class VersionData():
     """
-    This abstract class holds two types of data:
+    This class holds two types of data:
         - Data with a 1-to-1 relationship with the trial/version pair.
         - Data with a child relationship with the trial/version pair.
 
-    Each subclass will return the 1-to-1 data for another system to add to the DB.
+    This initializes with None attributes, and implements setter 
-    This is so that a single record can be created in one go.
+    methods to load them (just to double check types)
-    Each subclass will load the child data to the database directly. 
+    That way I can just pass around the VersionData instance
+    and add data as I go.
+
+    It will also implement the ability to load the data to the database
     """
-    @abc.abstractmethod
+    def __init__(self,nct_id,version_id):
-    def version_fields(self):
+        #identifiers
+        self.nct_id = nct_id
+        self.version_id = version_id
+
+        #Study Status
+        self._primary_completion_date = None
+        self._primary_completion_date_category = None
+        self._completion_date = None
+        self._completion_date_category = None
+        self._overall_status = None
+
+        #Study Design
+        self._enrollment = None
+        self._enrollment_category = None
+
+        #Sponsors and Collaborators
+        self._sponsor = None
+        #self._sponsor_category = None #I don't believe this is included in the raw data
+        self._responsible_party = None
+        #self._responsible_party_category = None #I don't believe this is included in the raw data
+        #self._collaborators = None #currently going to ignore as I've not fount it in AACT
+
+
+def extract_study_statuses(study_status_form, version_a,version_b):
     """
-        This function returns data that should be included in a standard table
+    This extracts data from a study_status form and returns one or two 
-        related to version_x of the record.
+    StudyStatusData objects, 
 
-        It also returns the columns?
     """
-        pass
+    #get rows
-    @abc.abstractmethod
+    rows = study_status_form.table.tbody.find_all("tr")
-    def version_records(self, foreign_key, db_cursor):
+    #iterate through rows, 
+    for trow in rows:
+        #matching on rowLabels
+        #print(trow.__str__()[:80])
+        match tr_to_td(trow):
+            case ["Primary Completion:" as row_label, tag]: 
+                old,new = split_by_version(tag)
+                tagdate1 = extract_date_and_tag(old.text,date_MMMM_YYYY)
+                version_a._primary_completion_date = tagdate1.date
+                version_a._primary_completion_date_category = tagdate1.tag
+                
+                tagdate2 = extract_date_and_tag(new.text,date_MMMM_YYYY)
+                version_b._primary_completion_date = tagdate2.date
+                version_b._primary_completion_date_category = tagdate2.tag
+
+            case ["Study Completion:" as row_label, tag]:
+                old,new = split_by_version(tag)
+                tagdate1 = extract_date_and_tag(old.text,date_MMMM_YYYY)
+                version_a._completion_date = tagdate1.date
+                version_a._completion_date_category = tagdate1.tag                
+
+                tagdate2 = extract_date_and_tag(new.text,date_MMMM_YYYY)
+                version_b._completion_date = tagdate2.date
+                version_b._completion_date_category = tagdate2.tag
+
+            case ["Overall Status:" as row_label, tag]:
+                old,new = split_by_version(tag)
+                version_a._overall_status = old.text
+                version_b._overall_status = new.text              
+
+
+def extract_study_design(study_status_form, version_a,version_b):
     """
-        This function loads data that needs to be held in auxilary tables
+    This extracts data from a study_status form and returns one or two 
-        into the database.
+    StudyStatusData objects, 
-        For example, the list of sponsors will need to be tracked separatly from 
+
-        trial status.
     """
-        pass
+    #get rows
+    rows = study_status_form.table.tbody.find_all("tr")
+    #iterate through rows, 
+    for trow in rows:
+        #matching on rowLabels
+        #print(trow.__str__()[:80])
+        match tr_to_td(trow):
+            case ["Enrollment:" as row_label, tag]: 
+                old,new = split_by_version(tag)
+                tagdate1 = extract_text_and_tag(old.text)
+                version_a._enrollment = tagdate1.text
+                version_a._enrollment_category = tagdate1.tag
                 
+                tagdate2 = extract_text_and_tag(new.text)
+                version_b._enrollment = tagdate2.text
+                version_b._enrollment_category = tagdate2.tag
 
-class StudyStatusData(VersionData):
-    columns = ["primary_completion_date", "completion_date", "last_update_posted_date"]
              
-    def __init__(self ,primary_completion_date, completion_date, last_update_posted_date) -> None:
+def extract_sponsor_data(study_status_form, version_a,version_b):
+    """
+    This extracts data from a study_status form and returns one or two 
+    StudyStatusData objects, 
+
+    """
+    #get rows
+    rows = study_status_form.table.tbody.find_all("tr")
+    #iterate through rows, 
+    for trow in rows:
+        #matching on rowLabels
+        #print(trow.__str__()[:80])
+        match tr_to_td(trow):
+            case ["Sponsor:" as row_label, tag]: 
+                old, new = split_by_version(tag)
+                version_a._sponsor = old.text
+                version_b._sponsor = new.text
+
+            case ["Responsible Party:" as row_label, tag]: 
+                old, new = split_by_version(tag)
+                version_a._responsible_party = old.text
+                version_b._responsible_party = new.text
+
+            case ["Collaborators:" as row_label, tag]: 
+                #old, new = split_by_version(tag)
+                #TODO: find a trial with multiple collaborators and figure out how to identify/count them:w
+                # So far can't figure out where this is in AACT, so I'm going to ignore it.
                 pass
 
-def extract_study_statuses(study_status_form, version_a,version_b):
+
+
+def split_by_version(tag):
+    #clone elements and remove sub-tags that are not needed.
+    old = copy(tag)
+    for span in old.find_all(class_="add_hilite"):
+        span.extract()
+        
+    new = copy(tag)
+    for span in new.find_all(class_="drop_hilite"):
+        span.extract()
+    return old,new
+
+
+def extract_date_and_tag(text, date_format):
     """
-    This extracts data from a study_status form and returns one or two StudyStatusData objects
+    Extracts a datetype according to the date format
+    and the estimate tag based on 
+
     """
-    pass
+    #FIX: Currently, there are multiple (mixed) data formats in use
+    #Theses can exist in the same data field, in two different versions
+    #so instead of using a single (passed) data format, I need to 
+    #select between various data formats.
 
-class SponsorCollaboratorsData(VersionData):
+    text = text.strip()
-    columns=[]
+
-    def __init__(self) -> None:
+    #handle various empty cases
-        pass
+    if not text or text == '':
+        return TagDatePair(None, None)
+        
+    date_split = text.split("[")
+    if len(date_split) > 1:
+        estimate_tag = date_split[1].split("]")[0].strip()
+    else:
+        estimate_tag = None
+    date_object = datetime.strptime(date_split[0].strip(), date_format)
+    
+    return TagDatePair(estimate_tag, date_object)
+
+
+def extract_text_and_tag(text):
+    """
+    Extracts a datetype according to the date format
+    and the estimate tag based on 
+
+    """
+    text = text.strip()
+
+    #handle various empty cases
+    if not text or text == '':
+        return TagDatePair(None, None)
         
+    date_split = text.split("[")
+    if len(date_split) > 1:
+        estimate_tag = date_split[1].split("]")[0].strip()
+    else:
+        estimate_tag = None
+    text_object = date_split[0].strip()
     
+    return TagTextPair(estimate_tag, text_object)
 
-def get_forms(soup):
+### FUNCTIONS
+
+def tr_to_td(tr) -> tuple[str, str]:
+    """
+    Takes an html data row of interest, extracts the record_name from the first <td>, and the data from the second <td>.
+
+    For the data, it just extracts the text. 
+    The text itself then needs processed separately, based on what it should contain.
+    """
+    #get list of cells
+    td_list = tr.find_all("td")
+    if len(td_list) == 2:
+        return td_list[0].text, td_list[1]
+    else:
+        return None, None
 
-    data_list = []
+def get_forms(soup,version_a,version_b):
 
     #extract all forms
     for form in soup.body.find_all("form"):
@@ -83,9 +231,9 @@ def get_forms(soup):
 
         match form.attrs["id"]:
             case "form_StudyStatus":
-                print("test successful 2")
+                extract_study_statuses(form,version_a,version_b)
             case "form_SponsorCollaborators":
-                pass
+                extract_sponsor_data(form, version_a, version_b)
             case "form_Oversight":
                 pass
             case "form_StudyDescription":
@@ -93,7 +241,7 @@ def get_forms(soup):
             case "form_Conditions":
                 pass
             case "form_StudyDesign":
-                pass
+                extract_study_design(form,version_a,version_b)
             case "form_ArmsandInterventions":
                 pass
             case "form_ProtocolOutcomeMeasures":
@@ -121,7 +269,19 @@ def get_forms(soup):
             case _:
                 print(form.attrs["id"])
 
+
+### CONSTANTS
+date_MMMM_YYYY = "%B %Y"
+date_MMMM_DD_YYYY = "%B %d, %Y"
+
 if __name__ == "__main__":
-    with open("./NCT00658567.html") as fh:
+
+    for file in ["./NCT00658567.html", "./NCT01303796.html"]:
+        with open(file) as fh:
             soup = BeautifulSoup(fh, "lxml")
-    get_forms(soup)
+        
+        version1 = VersionData("NCT00658567",1)
+        version2 = VersionData("NCT00658567",2)
+        get_forms(soup, version1, version2)
+        print(version1.__dict__) #order messed up somewhere:w
+        print(version2.__dict__) #order messed up somewhere:w

Parser/prototype_history.sql

@@ -6,9 +6,9 @@ CREATE TABLE history.versions
     nct_id
     version
     --Study Status
-    overall_status
+    overall_status^
-    primary_completion_date
+    primary_completion_date^
-    completion_date
+    completion_date^
     last_update_submitted_date
     --SponsorCollaborators
     sponsor (multi?)
@@ -31,7 +31,7 @@ CREATE TABLE history.versions
     Number of Arms
     Masking
     Allocation
-    Enrollment
+    Enrollment  ^
     --ArmsAndInterventions
     Arms (multiple) (Ignore)
     --ProtocolOutcomeMeasures

Parser/textprocessing.py

@@ -1,4 +1,6 @@
-from cgitb import html
+from copy import copy
+from datetime import datetime
+from bs4 import BeautifulSoup
 import re
 
 form = """
@@ -106,16 +108,77 @@ entry1 = """
 """
 
 
-drop_old_re = re.compile('<span class="drop_hilite">\w*</span>\s?')
+entry2 = '<td> <span class="add_hilite">December 2009 [Actual]</span> </td>'
-drop_new_re = re.compile('<span class="add_hilite">\w*</span>\s?')
-drop_tags_re = re.compile('<[=-_,.:;"/\w\s]+>')
 
+DROP_HILITE_re = re.compile('<span class="drop_hilite">[\[\]\w]*</span>\s?')
+ADD_HILITE_re = re.compile('<span class="add_hilite">\w*</span>\s?')
+TAGS_RE = re.compile('<[=-_,.:;"/\w\s]+>')
 
-print(drop_new_re.sub("",entry1))
+def extract_new_data(td):
-print(drop_old_re.sub("",entry1))
+    text = td.__str__()
-print(drop_tags_re.sub("",entry1))
+    return TAGS_RE.sub("",DROP_HILITE_re.sub(" ",text)).strip()
 
-print(drop_tags_re.sub("",drop_new_re.sub("",entry1)))
+def extract_old_data(td):
+    text = td.__str__()
+    return TAGS_RE.sub("",ADD_HILITE_re.sub(" ",text)).strip()
 
+def delete_tags(td):
+    text = td.__str__()
+    return TAGS_RE.sub(" ",text).strip()
 
-print(drop_tags_re.sub("",drop_new_re.sub("",form)))
+
+def extract_date_and_tag(text, date_format):
+    """
+    Extracts a datetype according to the date format
+    and the estimate tag based on 
+
+    """
+    if not text:
+        return " "
+        
+    date_split = text.split("[")
+    if len(date_split) > 1:
+        estimate_tag = date_split[1].split("]")[0].strip()
+    else:
+        estimate_tag = None
+    date_object = datetime.strptime(date_split[0].strip(), date_format)
+    
+    return estimate_tag, date_object
+    #TODO: Write test
+
+def extract_text_and_tag(text):
+    """
+    
+    """
+    pass
+
+if __name__ == "__main__":
+    Entry = BeautifulSoup(entry1, "lxml")
+    Form = BeautifulSoup(form, "lxml")
+
+
+
+    print(extract_new_data(Entry.find_all("td")[1]))
+    print(extract_old_data(Entry.find_all("td")[1]))
+
+    for tr in Form.find_all("tr"):
+        data = tr.find_all("td")
+        match len(data):
+            case 0: print("no data")
+            case 1: print("1\t",data[0])
+            case _: print(len(data), "\t", extract_new_data(data[1]) ,"\t|\t", extract_old_data(data[1]))
+
+    #print(extract_date_and_tag(extract_old_data(Entry.find_all("td")[1]), "%B %Y"))
+    print(extract_date_and_tag("April 2008 [ test ]", "%B %Y"))
+
+
+    Entry2 = BeautifulSoup(entry2,"lxml")
+    print(extract_old_data(Entry2)) #error here.
+    print(extract_new_data(Entry2))
+
+
+    Entry3 = copy(Entry2)
+    print(Entry3)
+    Entry4 = Entry3.find_all(class_="add_hilite")[0].extract()
+    print(Entry3.text)
+    print(Entry4.text)