Merged in bugfix/exhibit-header-detection (pull request #518)

Bugfix/exhibit header detection

* set up testing

* Increased context length, changed prompt to investment_prompts from preprocessing_prompts

* extracting the header

* minor changes

* Merged main into bugfix/exhibit-header-detection

* cleanup prints and tests


Approved-by: Alex Galarce

fieldExtraction/src/preprocessing_funcs.py

@@ -1,6 +1,6 @@
 import re
 
-from src.prompts import preprocessing_prompts
+from src.prompts import investment_prompts
 import src.utils.llm_utils as llm_utils
 import src.utils.string_utils as string_utils
 from src import config, keywords
@@ -225,9 +225,9 @@ def get_exhibit_pages(text_dict: dict[str, str], filename: str) -> list[str]:
             exhibit_pages.append(page_num)
             first_page = False
         else:
-            prompt = preprocessing_prompts.EXHIBIT_CHECK(page[0:100])
+            prompt = investment_prompts.EXHIBIT_CHECK(page[0:200])
             claude_answer_raw = llm_utils.invoke_claude(
-                prompt, config.MODEL_ID_CLAUDE3_HAIKU, filename, max_tokens=10 # TODO: low priority, try increasing max_tokens and maybe pass multiple pages in to reduce overall calls
+                prompt, config.MODEL_ID_CLAUDE3_HAIKU, filename, max_tokens=100 # TODO: low priority, try increasing max_tokens and maybe pass multiple pages in to reduce overall calls
             )
             claude_answer_extracted = string_utils.extract_text_from_delimiters(
                                 claude_answer_raw, Delimiter.PIPE