fixed the bug with function calling

.github/workflows/python-ci.yaml

@@ -12,7 +12,7 @@ jobs:
 
     strategy:
       matrix:
-        python-version: ["3.10", "3.11"]
+        python-version: ["3.11"]
 
     env:
       GROQ_API_KEY: ${{ secrets.GROQ_API_KEY }}

.gitignore

@@ -158,3 +158,8 @@ cython_debug/
 #  and can be added to the global gitignore or merged into this file.  For a more nuclear
 #  option (not recommended) you can uncomment the following to ignore the entire idea folder.
 .idea/
+
+.DS_Store
+
+# VSCode
+.vscode/

.vscode/launch.json

@@ -0,0 +1,15 @@
+{
+    // Use IntelliSense to learn about possible attributes.
+    // Hover to view descriptions of existing attributes.
+    // For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
+    "version": "0.2.0",
+    "configurations": [
+        {
+            "name": "Python Debugger: Current File",
+            "type": "debugpy",
+            "request": "launch",
+            "program": "${file}",
+            "console": "integratedTerminal"
+        }
+    ]
+}

environment.yaml

@@ -10,7 +10,7 @@ dependencies:
   - typer>=0.12.5 #for CLI
   - twine #for publishing
   - pip:
-      - litellm>=1.48.11
+      - litellm>=1.49.5
       - numpydoc
       - semanticscholar>=0.8.4
       - Mako>=1.3.5 #optional, might be deleted in the future

examples/function_calling.py

@@ -29,19 +29,28 @@ async def process_stream(async_generator):
         # You can also process each item here if needed
     return collected_data
 
-llm_options = just_agents.llm_options.OPENAI_GPT4oMINI
+#llm_options = just_agents.llm_options.OPENAI_GPT4oMINI
+llm_options = just_agents.llm_options.LLAMA3_2
+
 key_getter = rotate_env_keys
 prompt = "What's the weather like in San Francisco, Tokyo, and Paris?"
 
-
+load_dotenv(override=True)
 session: LLMSession = LLMSession(
-    llm_options=llm_options,
+    llm_options=just_agents.llm_options.OPENAI_GPT4oMINI,
     tools=[get_current_weather]
 )
+result = session.query("What's the weather like in San Francisco, Tokyo, and Paris?")
+
+
 session.memory.add_on_message(lambda m: pprint.pprint(m) if "content" in m is not None else None)
-session.query(prompt)
+result = session.query(prompt)
 
+"""
 print("And now same query but async mode for streaming. Note: we use asyncio.run here to run the stream")
 stream = session.stream(prompt)
 result = asyncio.run(process_stream(stream))
-print("stream finished")
+print("stream finished")
+"""
+print("RESULT+++++++++++++++++++++++++++++++++++++++++++++++")
+print(result)

just_agents/llm_session.py

@@ -36,17 +36,19 @@
 
 
 class LLMSession(IAgent):
-    available_tools: dict[str, Callable] = dict()
-    memory: Memory = Memory()
-    streaming: AbstractStreaming = None
-    key_getter: RotateKeys = None
-    on_response: list[OnCompletion] = []
+
 
 
     def __init__(self, llm_options: dict[str, Any] = None,
                  system_prompt:str = None,
                  agent_schema: str | Path | dict | None = None,
-                 tools: list[Callable] = None):
+                 tools: Optional[list[Callable]] = None):
+        self.on_response = []
+        self.available_tools: Optional[dict[str, Callable]] = {}
+        self.memory: Memory = Memory()
+        self.streaming: AbstractStreaming = None
+        self.key_getter: RotateKeys = None
+        self.on_response: list[OnCompletion] = []
 
         self.agent_schema = resolve_agent_schema(agent_schema, "LLMSession", "llm_session_schema.yaml")
         self.llm_options: dict[str, Any] = resolve_llm_options(self.agent_schema, llm_options)

setup.py

@@ -7,7 +7,7 @@
 with codecs.open(os.path.join(here, "README.md"), encoding="utf-8") as fh:
     long_description = "\n" + fh.read()
 
-VERSION = '0.1.1'
+VERSION = '0.2.0'
 DESCRIPTION = 'Just Agents'
 LONG_DESCRIPTION = 'LLM Agents that are implemented without unnecessary complexity'
 

tests/test_cot.py

@@ -1,6 +1,7 @@
 import asyncio
 from dotenv import load_dotenv
 from just_agents.cot_agent import ChainOfThoughtAgent
+import pytest
 
 import just_agents.llm_options
 
@@ -13,6 +14,7 @@ def count_letters(character:str, word:str) -> str:
     print("Function: ", character, " occurres in ", word, " ", count, " times.")
     return str(count)
 
+
 def test_function_query():
     load_dotenv(override = True)
 
@@ -27,6 +29,7 @@ async def process_stream(async_generator):
     for item in async_generator:
         print(item)
 
+
 def test_stream_function_query():
     load_dotenv(override = True)
 

tests/test_session.py

@@ -8,7 +8,7 @@
 import just_agents.llm_options
 from just_agents.llm_session import LLMSession
 import asyncio
-from mock import *
+from tests.mock import *
 
 
 def get_current_weather(location: str):
@@ -25,7 +25,7 @@ def get_current_weather(location: str):
 def test_sync_function_calling():
     load_dotenv(override=True)
     session: LLMSession = LLMSession(
-        llm_options=just_agents.llm_options.OPENAI_GPT4oMINI,
+        llm_options=just_agents.llm_options.LLAMA3_2,
         tools=[get_current_weather]
     )
     result = session.query("What's the weather like in San Francisco, Tokyo, and Paris?")
@@ -40,7 +40,7 @@ async def process_stream(async_generator):
 def test_stream_function_calling():
     load_dotenv(override=True)
     session: LLMSession = LLMSession(
-        llm_options=just_agents.llm_options.OPENAI_GPT4oMINI,
+        llm_options=just_agents.llm_options.LLAMA3_2,
         tools=[get_current_weather]
     )
     stream = session.stream("What's the weather like in San Francisco, Tokyo, and Paris?")
@@ -53,9 +53,9 @@ def test_stream_function_calling():
 
 
 def test_stream_genetics_function_calling():
-    load_dotenv()
+    load_dotenv(override=True)
     session: LLMSession = LLMSession(
-        llm_options=just_agents.llm_options.OPENAI_GPT4oMINI,
+        llm_options=just_agents.llm_options.LLAMA3_2,
         tools=[hybrid_search, rsid_lookup, gene_lookup, pathway_lookup, disease_lookup, sequencing_info, clinical_trails_full_trial]
     )
     stream = session.stream("What is the influence of different alleles in rs10937739?")