Merge pull request #4 from EJOOSTEROP/feat-cost_estimate

CHANGELOG.md

@@ -9,6 +9,14 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 
 - Consider downgrading Python requirement (at least to 3.10, but I dont know how low it can gowhere)
 
+## [0.8.4]
+
+### Added
+- Show actual cost - as calculated by langchain - after call. Note: does not seem to work for gpt-4o.
+
+### Changed
+- Catch error and show warning in log when command is run without OPENAI_API_KEY being set.
+
 ## [0.8.3]
 
 ### Fixed

pyproject.toml

@@ -18,7 +18,7 @@ disallow_incomplete_defs = true
 
 [tool.poetry]
 name = "crewcal"
-version = "0.8.3"
+version = "0.8.4"
 description = "Convert an airline crew schedule pdf into iCalendar format."
 authors = ["Erik Oosterop <[email protected]>"]
 license = "MIT"

src/crewcal/llm_extract.py

@@ -5,6 +5,7 @@
 
 About 0.75 US cents per call in OpenAI API costs.
 """
+
 import json
 import logging
 import os
@@ -22,7 +23,10 @@
 from crewcal.schedule import Schedule
 
 _ = load_dotenv(find_dotenv())
-openai.api_key = os.environ["OPENAI_API_KEY"]
+try:
+    openai.api_key = os.environ["OPENAI_API_KEY"]
+except KeyError:
+    logging.warning("WARNING - Environment variable OPENAI_API_KEY should be set.")
 
 
 class OpenAISchedule:
@@ -33,6 +37,7 @@ class OpenAISchedule:
 
     schedule_path: str
     extracted_schedule: str = ""
+    llm_model_name: str = "gpt-3.5-turbo"
 
     def __init__(
         self, schedule_path: str, to_json_file: str = "", to_icalendar_file: str = ""
@@ -85,7 +90,8 @@ def extract(self, to_file: str = "") -> None:
         full_sched_doc = self.read_schedule_pdf(self.schedule_path)
 
         if full_sched_doc:
-            model = ChatOpenAI(model_name="gpt-3.5-turbo", temperature=0)
+            # TODO: Can use gpt-4o instead; but app. 10x as expensive
+            model = ChatOpenAI(model_name=self.llm_model_name, temperature=0)
             prompt = ChatPromptTemplate.from_messages(
                 [("system", template_flight_schedule), ("human", "{input}")]
             )
@@ -106,10 +112,17 @@ def extract(self, to_file: str = "") -> None:
             )
 
             logging.warning(
-                "WARNING - This script costs 0.75 US cents per call in OpenAI API costs."
+                "WARNING - This script costs ~0.75 US cents per call in OpenAI API costs (GPT-3.5)."
             )
 
-            self.extracted_schedule = extraction_chain.invoke({"input": full_sched_doc})
+            from langchain.callbacks import get_openai_callback
+
+            with get_openai_callback() as cb:
+                self.extracted_schedule = extraction_chain.invoke(
+                    {"input": full_sched_doc}
+                )
+
+            logging.warning("Actual OpenAI API cost in USD:" + str(cb.total_cost))
 
         if to_file:
             self.write_json(to_file)