Update app.py

app.py

@@ -3,10 +3,12 @@ import re
 
 from datetime import datetime
 
+
 # load the spacy model
 spacy.cli.download("en_core_web_lg")
 nlp = spacy.load("en_core_web_lg")
 
+
 # Define a function to extract dates from text
 def extract_dates(text):
     """
@@ -17,10 +19,11 @@ def extract_dates(text):
     # Regular expressions that include the \b word boundary character to ensure that the date pattern only matches if it is not part of a longer pattern that has already been matched
     date_patterns = [
         r'\b\d{1,2}[-/]\d{1,2}[-/]\d{2,4}\b',  # Matches dates like "01/01/22" or "1-1-2022"
-        r'\b\d{1,2}[-/]\d{1,2}\b(?!\d)',       # Matches dates like "01/01" or "1-1", but not "10-10-2022"
+        r'\b\d{1,2}[-/]\d{1,2}\b(?!\d)',       # Matches dates like "01/01" or "1-1"
         r'\b[A-Z][a-z]{2,8} \d{1,2},? \d{2,4}\b',  # Matches dates like "January 1, 2022" or "Feb 28, 22"
         r'\b\d{1,2} [A-Z][a-z]{2,8} \d{2,4}\b',  # Matches dates like "1 January 2022" or "28 Feb 22"
         r'\b[A-Z][a-z]{2,8} \d{2,4}\b',        # Matches dates like "January 2022" or "Feb 22"
+        
     ]
 
     # Find all matches for date patterns in the text
@@ -48,88 +51,128 @@ def extract_dates(text):
 
 
 def convert_dates(date_list):
-    """
-    Assign to the identified formatted dates the proper date format and then, on the formatted dates, assign the relevant date tags (e.g. specify which is the day, the month, etc)
-    """
-    
-    output_list = []
-    for date_str in date_list:
-        try:
-            dt = datetime.strptime(date_str, '%B %d, %Y')
-            output_list.append(f"day:{dt.day}, month:{dt.month}, year:{dt.year}")
-        except ValueError:
-    
-            try:
-                dt = datetime.strptime(date_str, '%m-%d-%y')
-                output_list.append(f"day:{dt.day}, month:{dt.month}, year:{dt.year}")
-            except ValueError:
-    
-                try:
-                    dt = datetime.strptime(date_str, '%d/%m')
-                    output_list.append(f"day:{dt.day}, month:{dt.month}")
-                except ValueError:
-    
-                    try:
-                        dt = datetime.strptime(date_str, '%B %d')
-                        output_list.append(f"day:{dt.day}, month:{dt.month}")
-                    except ValueError:
-    
-                        try:
-                            dt = datetime.strptime(date_str, '%b %d')
-                            output_list.append(f"day:{dt.day}, month:{dt.month}")
-                        except ValueError:
-    
-                            try:
-                                dt = datetime.strptime(date_str, '%B %Y')
-                                output_list.append(f"month:{dt.month}, year:{dt.year}")
-                            except ValueError:
-    
-                                try:
-                                    dt = datetime.strptime(date_str, '%Y')
-                                    output_list.append(f"year:{dt.year}")
-                                except ValueError:
-                                    
-                                    try:
-                                        dt = datetime.strptime(date_str, '%d/%m/%y')
-                                        output_list.append(f"day:{dt.day}, month:{dt.month}, year:{dt.year}")
-                                    except ValueError:
-                                    
-                                      try:
-                                        dt = datetime.strptime(date_str, '%Y %d %m')
-                                        output_list.append({'day': dt.day, 'month': dt.month, 'year': dt.year})
-                                      except ValueError:
-                                    
-                                        try:
-                                            dt = datetime.strptime(date_str, '%B')
-                                            output_list.append(dt.strftime('month:%-m'))
-                                        except ValueError as e:
-                                            output_list.append(f'INVALID FORMAT: {date_str}')
-    return output_list
+  """
+  Assign to the identified formatted dates the proper date format and then, on the formatted dates, assign the relevant date tags (e.g. specify which is the day, the month, etc)
+  """
+
+  DATE_FORMATS = {
+      '%B %d, %Y': 'day:{dt.day}, month:{dt.month}, year:{dt.year}',
+      '%-m-%d-%Y': 'day:{dt.day}, month:{dt.month}, year:{dt.year}',
+      '%m-%d-%y': 'day:{dt.day}, month:{dt.month}, year:{dt.year}',
+      '%d/%m': 'day:{dt.day}, month:{dt.month}',
+      '%B %d': 'day:{dt.day}, month:{dt.month}',
+      '%b %d': 'day:{dt.day}, month:{dt.month}',
+      '%B %Y': 'month:{dt.month}, year:{dt.year}',
+      '%Y': 'year:{dt.year}',
+      '%d/%m/%y': 'day:{dt.day}, month:{dt.month}, year:{dt.year}',
+      '%B %d, %y': 'day:{dt.day}, month:{dt.month}, year:{dt.year}',
+      '%b %d, %y': 'day:{dt.day}, month:{dt.month}, year:{dt.year}',
+      '%d-%m-%Y': 'day:{dt.day}, month:{dt.month}, year:{dt.year}',
+      '%d/%m/%Y': 'day:{dt.day}, month:{dt.month}, year:{dt.year}',
+      '%d-%m-%y': 'day:{dt.day}, month:{dt.month}, year:{dt.year}',
+      '%m/%d/%y': 'day:{dt.day}, month:{dt.month}, year:{dt.year}',
+      '%m/%d/%Y': 'day:{dt.day}, month:{dt.month}, year:{dt.year}',
+      '%m-%d-%Y': 'day:{dt.day}, month:{dt.month}, year:{dt.year}',
+      '%m-%d-%y': 'day:{dt.day}, month:{dt.month}, year:{dt.year}',
+      '%d/%m/%Y %H:%M:%S': 'day:{dt.day}, month:{dt.month}, year:{dt.year}, time:{dt.strftime("%H:%M:%S")}',
+      '%d/%m/%y %H:%M:%S': 'day:{dt.day}, month:{dt.month}, year:{dt.year}, time:{dt.strftime("%H:%M:%S")}',
+      '%m/%d/%Y %H:%M:%S': 'day:{dt.day}, month:{dt.month}, year:{dt.year}, time:{dt.strftime("%H:%M:%S")}',
+      '%m/%d/%y %H:%M:%S': 'day:{dt.day}, month:{dt.month}, year:{dt.year}, time:{dt.strftime("%H:%M:%S")}',
+      '%Y-%m-%d': 'day:{dt.day}, month:{dt.month}, year:{dt.year}',
+      '%y-%m-%d': 'day:{dt.day}, month:{dt.month}, year:{dt.year}',
+      '%m-%d-%Y %H:%M:%S': 'day:{dt.day}, month:{dt.month}, year:{dt.year}, time:{dt.strftime("%H:%M:%S")}',
+      '%m-%d-%y %H:%M:%S': 'day:{dt.day}, month:{dt.month}, year:{dt.year}, time:{dt.strftime("%H:%M:%S")}',
+      '%m-%d': 'month:{dt.month}, day:{dt.day}',
+      '%-m-%-d': 'month:{dt.month}, day:{dt.day}',
+      '%d %b %y': 'day:{dt.day}, month:{dt.month}, year:{dt.year}',
+      '%d %B %Y': 'day:{dt.day}, month:{dt.month}, year:{dt.year}',
+      '%b %Y': 'month:{dt.month}, year:{dt.year}',
+      '%b %d, %Y': 'day:{dt.day}, month:{dt.month}, year:{dt.year}',
+      '%d %B %y': 'day:{dt.day}, month:{dt.month}, year:{dt.year}'  
+  }
+
+
+  output_list = []
+  for date_str in date_list:
+      valid_format = False
+      for fmt, out_fmt in DATE_FORMATS.items():
+          try:
+              dt = datetime.strptime(date_str, fmt)
+              output_list.append(out_fmt.format(dt=dt))
+              valid_format = True
+              break
+          except ValueError:
+              pass
+      if not valid_format:
+          # Attempt to parse using a custom format
+          try:
+              if '-' in date_str:
+                  dt = datetime.strptime(date_str, '%m-%d-%y')
+              else:
+                  dt = datetime.strptime(date_str, '%d/%m/%y')
+              output_list.append(f'day:{dt.day}, month:{dt.month}, year:{dt.year}')
+          except ValueError:
+              output_list.append(f'INVALID FORMAT: {date_str}')
+  return output_list
 
 
 
 def dates_binding(text):
+  '''
+  This is a function that binds together all the subcomponents of the dates identification, while also controlling for multiple, or zero date references
+  '''
+
+  try:
+
+      # capture the referred dates
+      identified_dates = extract_dates(text)
+
+      # we only accept for one date reference
+      if len(identified_dates) == 1:
+
+        formatted_dates = convert_dates(identified_dates)
+
+
+        # in case there is a wrong date format then return the appropriate code to prompt back the proper message
+        if 'INVALID FORMAT' in formatted_dates[0]:
+          return (0,'DATES','wrong_date_format')
+
+        else:
+          return formatted_dates
+
+      # in case of zero references return the appropriate code (to aid returning the correct prompt)
+      elif len(identified_dates) == 0:
+        return (0,'DATES','no_date')
+
+      # in case of more than one references return the appropriate code (to aid returning the correct prompt)
+      elif len(identified_dates) > 1:
+        return (0,'DATES','more_dates')
+
+      # in case of unexpected error return the appropriate code (to aid returning the correct prompt)
+      else:
+        return (0,'DATES','unknown_error')
 
- identified_dates = extract_dates(text)
- formatted_dates = convert_dates(identified_dates)
+  except:
+      return (0,'DATES','unknown_error')
 
- return formatted_dates
 
 
 from transformers import pipeline
 import gradio as gr
 
-title = "Natural Language module Demo for Dates"
-description = "This is a simple demo just for demonstration purposes, so that Serco team might have the chance to validate the results of the Natural Language module concerning date identification, while in progress"
+title = "Dates Demo"
+description = "This is a simple demo just for demonstration purposes for Serco team, to validate the results of the Natural Language module concerning dates identification, while in progress"
 
 examples = [
-    ["Earthquake located in Ishkoshim, Tajikistan in 2018 05 09 with magnitude greater than 6.2"],
-    ["Give me all the earthquakes with magnitude above than 6 in the region of Athens for the month of January 2, 2023"],
-    ["Earthquake happened in Rome during 1986 with a magnitude of 6"],
-    ["I want all earthquakes larger than 5.0 that occurred in Rome during 3/5/20"],
-    ["I want all earthquakes happened in Greece 31/12"],
-    ["I want all earthquakes happened in Greece January 5"],
-    ["I want all earthquakes happened in Greece 3-4-20"]
+    ["Earthquake located in Rome, Italy in 01/01/23 with magnitude greater than 6.2"],
+    ["Earthquake located in Rome, Italy in 1-1-2023 with magnitude greater than 6.2"],
+    ["Earthquake located in Rome, Italy in 1/1/23 with magnitude greater than 6.2"],
+    ["Earthquake located in Rome, Italy in January 1, 2023 with magnitude greater than 6.2"],
+    ["Earthquake located in Rome, Italy in Jan 1, 23 with magnitude greater than 6.2"],
+    ["Earthquake located in Rome, Italy in 1 January 2023 with magnitude greater than 6.2"],
+    ["Earthquake located in Rome, Italy in 1 Jan 23 with magnitude greater than 6.2"],
+    ["Earthquake located in Rome, Italy in January 2023 with magnitude greater than 6.2"],
+    ["Earthquake located in Rome, Italy in Jan 23 with magnitude greater than 6.2"],
 ]
 
 gr.Interface(