jacob-mathison/doczyai-pipelines
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Activity

resolve hsc issue

Browse Source
This commit is contained in:
Katon Minhas
2026-02-02 22:23:23 -05:00
parent 9b30c5f282
commit 62605981c1
3 changed files with 15 additions and 31 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/pipelines/shared/preprocessing/hybrid_smart_chunking_funcs.py
+5 -16
View File
@@ -279,14 +279,6 @@ def run_hybrid_smart_chunked_fields(
answers_dict, field_name="PROVIDER_STATE"
)
# Normalize PAYER_STATE to standardized two-letter abbreviation
for state_field in ["PAYER_STATE"]:
if state_field in answers_dict:
answers_dict[state_field] = (
string_utils.normalize_state_to_abbreviation(
answers_dict[state_field]
)
)
# Normalize PAYER_STATE
answers_dict = string_utils.normalize_state_field(
answers_dict, field_name="PAYER_STATE"
@@ -313,7 +305,7 @@ def extract_amendment_num_from_filename(answer_dict: dict, filename: str) -> dic
or answer_dict.get("CONTRACT_AMENDMENT_NUM") == "UNKNOWN"
):
prompt = prompt_templates.EXTRACT_AMENDMENT_NUM_FROM_FILENAME(filename)
prompt, _parser = prompt_templates.EXTRACT_AMENDMENT_NUM_FROM_FILENAME(filename)
llm_answer_raw = llm_utils.invoke_claude(
prompt,
"sonnet_latest",
@@ -324,7 +316,7 @@ def extract_amendment_num_from_filename(answer_dict: dict, filename: str) -> dic
)
try:
llm_answer_final = string_utils.universal_json_load(llm_answer_raw)
llm_answer_final = _parser(llm_answer_raw)
answer_dict["FILENAME_AMENDMENT_NUM"] = llm_answer_final.get(
"amendment_number", "N/A"
)
@@ -389,13 +381,10 @@ def prompt_hsc_single_field(
try:
llm_answer_final = _parser(llm_answer_raw) # returns dict
val = llm_answer_final.get(field.field_name)
if isinstance(val, list):
llm_answer_final[field.field_name] = val
elif isinstance(val, str):
llm_answer_final[field.field_name] = [val]
field_value = llm_answer_final.get(field.field_name)
if not isinstance(field_value, list):
field_value = [field_value]
field_value = llm_answer_final[field.field_name]
if field_value[0] == "N/A":
field.field_type = "full_context"
if field.field_name == "PAYER_NAME":