mixpanel analytics & config refactoring

app.py

@@ -2,7 +2,7 @@ import requests
 import zipfile
 import io
 from know_lang_bot.chat_bot.chat_interface import create_chatbot
-from know_lang_bot.config import AppConfig
+from know_lang_bot.configs.config import AppConfig
 import tempfile
 from rich.console import Console
 

poetry.lock

@@ -2157,6 +2157,24 @@ typing-inspect = ">=0.9.0"
 [package.extras]
 gcp = ["google-auth (>=2.27.0)", "requests (>=2.32.3)"]
 
+[[package]]
+name = "mixpanel"
+version = "4.10.1"
+description = "Official Mixpanel library for Python"
+optional = false
+python-versions = ">=2.7, !=3.4.*"
+groups = ["main"]
+markers = "python_version <= \"3.11\" or python_version >= \"3.12\""
+files = [
+    {file = "mixpanel-4.10.1-py2.py3-none-any.whl", hash = "sha256:a7a338b7197327e36356dbc1903086e7626db6d88367ccdd732b3f3c60d3b3ed"},
+    {file = "mixpanel-4.10.1.tar.gz", hash = "sha256:29a6b5773dd34f05cf8e249f4e1d16e7b6280d6b58894551ce9a5aad7700a115"},
+]
+
+[package.dependencies]
+requests = ">=2.4.2"
+six = ">=1.9.0"
+urllib3 = "*"
+
 [[package]]
 name = "mmh3"
 version = "5.1.0"
@@ -5329,4 +5347,4 @@ type = ["pytest-mypy"]
 [metadata]
 lock-version = "2.1"
 python-versions = ">=3.10, <3.13"
-content-hash = "4877a761765907ab7a7f19e8933a36d76289b82b9670fc2feeb9691db00bee39"
+content-hash = "9018994b12de342028b9f7a761de0f77f618a3e7489bebf2756e684fac5c00dd"

pyproject.toml

@@ -20,7 +20,8 @@ dependencies = [
     "chromadb (>=0.6.3,<0.7.0)",
     "ollama (>=0.4.7,<0.5.0)",
     "gradio (>=5.13.1,<6.0.0)",
-    "voyageai (>=0.3.2,<0.4.0)"
+    "voyageai (>=0.3.2,<0.4.0)",
+    "mixpanel (>=4.10.1,<5.0.0)"
 ]
 
 [tool.poetry]

src/know_lang_bot/__main__.py

@@ -7,7 +7,7 @@ from rich.console import Console
 from rich.table import Table
 
 from know_lang_bot.core.types import CodeChunk
-from know_lang_bot.config import AppConfig
+from know_lang_bot.configs.config import AppConfig
 from know_lang_bot.parser.factory import CodeParserFactory
 from know_lang_bot.parser.providers.git import GitProvider
 from know_lang_bot.parser.providers.filesystem import FilesystemProvider

src/know_lang_bot/chat_bot/__main__.py

@@ -1,4 +1,4 @@
-from know_lang_bot.config import AppConfig
+from know_lang_bot.configs.config import AppConfig
 from know_lang_bot.chat_bot.chat_graph import process_chat
 import chromadb
 import asyncio

src/know_lang_bot/chat_bot/chat_graph.py

@@ -5,7 +5,7 @@ from typing import AsyncGenerator, List, Dict, Any, Optional
 import chromadb
 from pydantic import BaseModel
 from pydantic_graph import BaseNode, EndStep, Graph, GraphRunContext, End, HistoryStep
-from know_lang_bot.config import AppConfig, RerankerConfig, EmbeddingConfig
+from know_lang_bot.configs.config import AppConfig, RerankerConfig, EmbeddingConfig
 from know_lang_bot.utils.fancy_log import FancyLogger
 from pydantic_ai import Agent
 import logfire

src/know_lang_bot/chat_bot/chat_interface.py

@@ -1,8 +1,10 @@
+from dataclasses import dataclass
 import gradio as gr
-from know_lang_bot.config import AppConfig
+from know_lang_bot.configs.config import AppConfig
 from know_lang_bot.utils.fancy_log import FancyLogger
 from know_lang_bot.utils.rate_limiter import RateLimiter
 from know_lang_bot.chat_bot.chat_graph import stream_chat_progress, ChatStatus
+from know_lang_bot.chat_bot.feedback import ChatAnalytics
 import chromadb
 from typing import List, Dict, AsyncGenerator
 from pathlib import Path
@@ -11,12 +13,54 @@ from gradio import ChatMessage
 
 LOG = FancyLogger(__name__)
 
+@dataclass
+class CodeContext:
+    file_path: str
+    start_line: int
+    end_line: int
+
+    def to_title(self) -> str:
+        """Format code context as a title string"""
+        title = f"📄 {self.file_path} (lines {self.start_line}-{self.end_line})"
+        return title
+    
+    @classmethod
+    def from_title(cls, title: str) -> "CodeContext":
+        """Parse code context from a title string"""
+        try:
+            # Split on first emoji and space
+            parts = title.split("📄 ")[1].split(" ")
+            file_path = parts[0]
+            # Extract line numbers from parentheses
+            line_nums = parts[1].strip("()").split(" ")[1].split("-")
+            start_line = int(line_nums[0])
+            end_line = int(line_nums[1])
+            
+            return cls(
+                file_path=file_path,
+                start_line=start_line,
+                end_line=end_line,
+            )
+        except Exception as e:
+            LOG.error(f"Error parsing code context from title: {e}")
+            return None
+
+    @classmethod
+    def from_metadata(cls, metadata: Dict) -> "CodeContext":
+        """Create code context from metadata dictionary"""
+        return cls(
+            file_path=metadata['file_path'],
+            start_line=metadata['start_line'],
+            end_line=metadata['end_line'],
+        )
+
 class CodeQAChatInterface:
     def __init__(self, config: AppConfig):
         self.config = config
         self._init_chroma()
         self.codebase_dir = Path(config.db.codebase_directory)
         self.rate_limiter = RateLimiter()
+        self.chat_analytics = ChatAnalytics(config.chat_analytics)
         
     def _init_chroma(self):
         """Initialize ChromaDB connection"""
@@ -41,20 +85,28 @@ class CodeQAChatInterface:
 
     def _format_code_block(self, metadata: Dict) -> str:
         """Format a single code block with metadata"""
-        file_path = metadata['file_path']
-        start_line = metadata['start_line']
-        end_line = metadata['end_line']
-        
-        code = self._get_code_block(file_path, start_line, end_line)
+        context = CodeContext.from_metadata(metadata)
+        code = self._get_code_block(
+            context.file_path, 
+            context.start_line, 
+            context.end_line
+        )
         if not code:
             return None
-            
-        title = f"📄 {file_path} (lines {start_line}-{end_line})"
-        if metadata.get('name'):
-            title += f" - {metadata['type']}: {metadata['name']}"
-            
 
-        return f"<details><summary>{title}</summary>\n\n```python\n{code}\n```\n\n</details>"
+        return f"<details><summary>{context.to_title()}</summary>\n\n```python\n{code}\n```\n\n</details>"
+    
+    def _handle_feedback(self, like_data: gr.LikeData, history: List[ChatMessage], request: gr.Request):
+         # Get the query and response pair
+        query = history[like_data.index - 1]["content"]  # User question
+        
+        # Track feedback
+        self.chat_analytics.track_feedback(
+            like=like_data.value,  # True for thumbs up, False for thumbs down
+            query=query,
+            client_ip=request.request.client.host
+        )
+
 
     async def stream_response(
         self,
@@ -178,12 +230,14 @@ class CodeQAChatInterface:
                 clear = gr.ClearButton([msg, chatbot], scale=1)
 
             async def respond(message: str, history: List[ChatMessage], request: gr.Request) -> AsyncGenerator[List[ChatMessage], None]:
+                self.chat_analytics.track_query(message, request.request.client.host)
                 async for updated_history in self.stream_response(message, history, request):
                     yield updated_history
                         
             # Set up event handlers
             msg.submit(respond, [msg, chatbot], [chatbot])
             submit.click(respond, [msg, chatbot], [chatbot])
+            chatbot.like(self._handle_feedback, [chatbot])
 
         return interface
 

src/know_lang_bot/chat_bot/feedback.py

@@ -0,0 +1,43 @@
+from mixpanel import Mixpanel
+from datetime import datetime
+from enum import Enum
+from know_lang_bot.configs.chat_config import ChatbotAnalyticsConfig
+
+class ChatFeedback(Enum):
+    POSITIVE = "positive"
+    NEGATIVE = "negative"
+
+
+class ChatAnalytics:
+    def __init__(self, config: ChatbotAnalyticsConfig):
+        self.mp = Mixpanel(config.api_key)
+        
+    def track_query(
+        self, 
+        query: str, 
+        client_ip: str
+    ):
+        """Track query event in Mixpanel"""
+        self.mp.track(
+            distinct_id=hash(client_ip),  # Hash for privacy
+            event_name="chat_query",
+            properties={
+                "query": query,
+            }
+        )
+
+    def track_feedback(
+        self, 
+        like: bool,
+        query: str,
+        client_ip: str
+    ):
+        """Track feedback event in Mixpanel"""
+        self.mp.track(
+            distinct_id=hash(client_ip),  # Hash for privacy
+            event_name="chat_feedback",
+            properties={
+                "feedback": ChatFeedback.POSITIVE.value if like else ChatFeedback.NEGATIVE.value,
+                "query": query,
+            }
+        )

src/know_lang_bot/chat_bot/gradio_demo.py

@@ -1,5 +1,5 @@
 from know_lang_bot.chat_bot.chat_interface import create_chatbot
-from know_lang_bot.config import AppConfig
+from know_lang_bot.configs.config import AppConfig
 
 
 config = AppConfig()

src/know_lang_bot/configs/chat_config.py

@@ -0,0 +1,23 @@
+from pydantic import Field
+from pydantic_settings import BaseSettings
+from enum import Enum
+
+
+class AnalyticsProvider(str, Enum):
+    MIXPANEL = "mixpanel"
+
+
+class ChatbotAnalyticsConfig(BaseSettings):
+    enabled: bool = Field(
+        default=False,
+        description="Enable analytics tracking"
+    )
+    provider: AnalyticsProvider = Field(
+        default=AnalyticsProvider.MIXPANEL,
+        description="Analytics provider to use for tracking feedback"
+    )
+
+    api_key: str = Field(
+        default=None,
+        description="api key for feedback tracking"
+    )

src/know_lang_bot/{config.py → configs/config.py}

@@ -4,6 +4,7 @@ from pydantic import Field, field_validator, ValidationInfo
 from pathlib import Path
 import fnmatch
 from know_lang_bot.core.types import ModelProvider
+from know_lang_bot.configs.chat_config import ChatbotAnalyticsConfig
 import os
 
 def _validate_api_key(v: Optional[str], info: ValidationInfo) -> Optional[str]:
@@ -215,4 +216,5 @@ class AppConfig(BaseSettings):
     db: DBConfig = Field(default_factory=DBConfig)
     parser: ParserConfig = Field(default_factory=ParserConfig)
     chat: ChatConfig = Field(default_factory=ChatConfig)
-    embedding: EmbeddingConfig = Field(default_factory=EmbeddingConfig)
+    embedding: EmbeddingConfig = Field(default_factory=EmbeddingConfig)
+    chat_analytics: ChatbotAnalyticsConfig = Field(default_factory=ChatbotAnalyticsConfig)

src/know_lang_bot/evaluation/chatbot_evaluation.py

@@ -2,7 +2,7 @@ from typing import List
 from enum import Enum
 from pydantic import BaseModel, Field, computed_field
 from pydantic_ai import Agent
-from know_lang_bot.config import AppConfig
+from know_lang_bot.configs.config import AppConfig
 from know_lang_bot.utils.chunking_util import truncate_chunk
 from know_lang_bot.utils.model_provider import create_pydantic_model
 from know_lang_bot.chat_bot.chat_graph import ChatResult, process_chat

src/know_lang_bot/evaluation/embedding_evaluation.py

@@ -7,7 +7,7 @@ import numpy as np
 from pydantic import BaseModel
 from dataclasses import dataclass
 from know_lang_bot.chat_bot.chat_graph import ChatResult
-from know_lang_bot.config import AppConfig, EmbeddingConfig
+from know_lang_bot.configs.config import AppConfig, EmbeddingConfig
 import json
 from know_lang_bot.evaluation.chatbot_evaluation import EvalCase, TRANSFORMER_TEST_CASES
 from know_lang_bot.models.embeddings import EmbeddingInputType, generate_embedding, EmbeddingVector

src/know_lang_bot/models/batch_request.py

@@ -1,5 +1,5 @@
 from typing import List
-from know_lang_bot.config import ModelProvider, LLMConfig
+from know_lang_bot.configs.config import ModelProvider, LLMConfig
 from pydantic_ai.messages import ModelMessage
 
 def _process_anthropic_batch(batched_input: List[List[ModelMessage]], config: LLMConfig) -> List[str]:

src/know_lang_bot/models/embeddings.py

@@ -2,7 +2,7 @@ import ollama
 import openai
 import voyageai
 import voyageai.client
-from know_lang_bot.config import EmbeddingConfig, ModelProvider
+from know_lang_bot.configs.config import EmbeddingConfig, ModelProvider
 from typing import Union, List, overload, Optional
 from enum import Enum
 

src/know_lang_bot/parser/base/parser.py

@@ -2,7 +2,7 @@ from abc import ABC, abstractmethod
 from typing import List
 from pathlib import Path
 from know_lang_bot.core.types import CodeChunk
-from know_lang_bot.config import AppConfig, LanguageConfig
+from know_lang_bot.configs.config import AppConfig, LanguageConfig
 from tree_sitter import Language, Parser
 
 class LanguageParser(ABC):

src/know_lang_bot/parser/base/provider.py

@@ -1,7 +1,7 @@
 from abc import ABC, abstractmethod
 from typing import Generator
 from pathlib import Path
-from know_lang_bot.config import AppConfig
+from know_lang_bot.configs.config import AppConfig
 
 class CodeProvider(ABC):
     """Abstract base class for code source providers"""

src/know_lang_bot/parser/factory.py

@@ -3,7 +3,7 @@ from pathlib import Path
 
 from know_lang_bot.parser.base.parser import LanguageParser
 from know_lang_bot.parser.languages.python.parser import PythonParser
-from know_lang_bot.config import AppConfig
+from know_lang_bot.configs.config import AppConfig
 
 class CodeParserFactory():
     """Concrete implementation of parser factory"""

src/know_lang_bot/parser/providers/filesystem.py

@@ -1,6 +1,6 @@
 from pathlib import Path
 from typing import Generator
-from know_lang_bot.config import AppConfig
+from know_lang_bot.configs.config import AppConfig
 from know_lang_bot.parser.base.provider import CodeProvider
 import os
 

src/know_lang_bot/parser/providers/git.py

@@ -1,7 +1,7 @@
 from pathlib import Path
 from typing import Generator
 from git import Repo
-from know_lang_bot.config import AppConfig
+from know_lang_bot.configs.config import AppConfig
 from know_lang_bot.parser.base.provider import CodeProvider
 import os
 

src/know_lang_bot/summarizer/summarizer.py

@@ -6,7 +6,7 @@ from pydantic import BaseModel, Field
 from pprint import pformat
 from rich.progress import Progress
 
-from know_lang_bot.config import AppConfig
+from know_lang_bot.configs.config import AppConfig
 from know_lang_bot.core.types import CodeChunk, ModelProvider
 from know_lang_bot.utils.fancy_log import FancyLogger
 from know_lang_bot.utils.model_provider import create_pydantic_model

src/know_lang_bot/utils/migration/openai_embedding_migrations.py

@@ -8,7 +8,7 @@ from rich.console import Console
 from typing import List, Dict, Optional
 from openai import OpenAI
 from datetime import datetime
-from know_lang_bot.config import AppConfig
+from know_lang_bot.configs.config import AppConfig
 from know_lang_bot.utils.fancy_log import FancyLogger
 from know_lang_bot.utils.chunking_util import truncate_chunk
 

src/know_lang_bot/utils/migration/voyage_embedding_migraions.py

@@ -5,7 +5,7 @@ from chromadb.errors import InvalidCollectionException
 from rich.progress import Progress
 from rich.console import Console
 from typing import List
-from know_lang_bot.config import AppConfig, EmbeddingConfig
+from know_lang_bot.configs.config import AppConfig, EmbeddingConfig
 from know_lang_bot.models.embeddings import generate_embedding, EmbeddingInputType
 from know_lang_bot.utils.fancy_log import FancyLogger
 

tests/conftest.py

@@ -3,7 +3,7 @@ import tempfile
 import git
 from pathlib import Path
 from typing import Dict
-from know_lang_bot.config import AppConfig, ParserConfig, LanguageConfig
+from know_lang_bot.configs.config import AppConfig, ParserConfig, LanguageConfig
 from know_lang_bot.parser.languages.python.parser import PythonParser
 from tests.test_data.python_files import TEST_FILES
 

tests/test_summarizer.py

@@ -4,7 +4,7 @@ from unittest.mock import Mock, patch, AsyncMock
 from pathlib import Path
 from know_lang_bot.summarizer.summarizer import CodeSummarizer
 from know_lang_bot.core.types import CodeChunk, ChunkType
-from know_lang_bot.config import AppConfig
+from know_lang_bot.configs.config import AppConfig
 
 @pytest.fixture
 def config():