added end date calculation based on pevious info

bptlab · Jan 16, 2024 · 9f06697 · 9f06697
1 parent c0ef120
commit 9f06697
Show file tree

Hide file tree

Showing 3 changed files with 6 additions and 4 deletions.
diff --git a/test.py b/test.py
@@ -11,11 +11,11 @@
 from tracex.extraction.prototype import function_calls as fc
 from tracex.extraction.prototype import prompts as p
 
-text = open(u.input_path / "journey_synth_covid_0.txt").read()
+text = open(u.input_path / "journey_synth_covid_1.txt").read()
 df = ih.convert_text_to_bulletpoints(text)
 print(df)
 df = ih.add_start_dates(text, df)
-# df = ih.add_end_dates(text, df)
+df = ih.add_end_dates(text, df)
 # df = ih.add_durations(text, df)
 # df = ih.add_event_types(df)
 # df = ih.add_locations(df)

diff --git a/tracex/extraction/prototype/input_inquiry.py b/tracex/extraction/prototype/input_inquiry.py
@@ -58,7 +58,7 @@ def create_patient_journey():
         {"role": "system", "content": p.create_patient_journey_context()},
         {"role": "user", "content": p.CREATE_PATIENT_JOURNEY_PROMPT},
     ]
-    patient_journey_txt = u.query_gpt(messages, u.TEMPERATURE_CREATION)
+    patient_journey_txt = u.query_gpt(messages=messages, temperature=u.TEMPERATURE_CREATION)
     i = 0
     proposed_filename = "journey_synth_covid_" + str(i) + ".txt"
     output_path = u.input_path / proposed_filename

diff --git a/tracex/extraction/prototype/prompts.py b/tracex/extraction/prototype/prompts.py
@@ -165,7 +165,9 @@ def life_circumstances_prompt(sex):
     Only return the date! Nothing else!
 """
 END_DATE_PROMPT = """
-    Here is the text and the bulletpoint with the start date for which you should extract end date in the format YYYYMMDD with the postfix T0000!
+    Here is the text and the bulletpoint with the start date for which you should extract the end date in the format YYYYMMDD with the postfix T0000! 
+    In case that you are not able to find a end date return the term "N/A". Only use the format YYYYMMDDTHHMM e.g. 20200401T0000!
+    Explain step by step your conclusions if the end date YYYYMMDDTHHMM is available, if not calculate the average time of the activity and add this on the start date resulting as the end date.
 """
 END_DATE_ANSWER = """
     For example for the text 'Four days after the first april 2020 I went to the doctor and got tested positive for Covid19. I was then hospitalized for two weeks.'