Merge pull request #77 from ansari-project/code-reorg

Moving files around to adhere to Python standards

.github/workflows/python-app.yml

@@ -28,7 +28,8 @@ jobs:
       WHATSAPP_BUSINESS_PHONE_NUMBER_ID: ${{ secrets.WHATSAPP_BUSINESS_PHONE_NUMBER_ID }}
       WHATSAPP_ACCESS_TOKEN_FROM_SYS_USER: ${{ secrets.WHATSAPP_ACCESS_TOKEN_FROM_SYS_USER }}
       WHATSAPP_VERIFY_TOKEN_FOR_WEBHOOK: ${{ secrets.WHATSAPP_VERIFY_TOKEN_FOR_WEBHOOK }}
-
+      PYTHONPATH: src
+
     container: python:3.10
     services:
       postgres:

pyproject.toml

@@ -0,0 +1,23 @@
+[build-system]
+requires = ["setuptools", "wheel"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "ansari-backend"
+version = "0.1.0"
+description = "Ansari is an AI assistant to enhance understanding and practice of Islam."
+authors = [
+    { name = "Ansari Project", email = "feedback@ansari.chat" }
+]
+requires-python = ">=3.8"
+
+[project.urls]
+Homepage = "https://github.com/ansari-project/ansari-backend"
+Documentation = "https://github.com/ansari-project/ansari-backend"
+Source = "https://github.com/ansari-project/ansari-backend"
+Tracker = "https://github.com/ansari-project/ansari-backend/issues"
+
+[tool.ruff]
+line-length = 88
+lint.select = ["E", "F", "W"]
+lint.ignore = ["E501"]

pytest.ini

@@ -0,0 +1,3 @@
+[pytest]
+markers =
+    asyncio: mark a test as asyncio

requirements.txt

@@ -1,4 +1,5 @@
 bcrypt
+build
 discord.py
 diskcache
 fastapi
@@ -14,10 +15,13 @@ psycopg2-binary
 pydantic_settings
 pyislam
 pyjwt
+pytest-asyncio
 rich
 sendgrid
+setuptools
 tenacity
 tiktoken
 typer
 uvicorn
+wheel
 zxcvbn

src/ansari/__init__.py

@@ -0,0 +1,4 @@
+# This file marks the directory as a Python package.
+from .config import Settings, get_settings
+
+__all__ = ["Settings", "get_settings"]

src/ansari/agents/__init__.py

@@ -0,0 +1,4 @@
+from .ansari import Ansari
+from .ansari_workflow import AnsariWorkflow
+
+__all__ = ["Ansari", "AnsariWorkflow"]

agents/ansari.py → src/ansari/agents/ansari.py

@@ -2,8 +2,6 @@
 import json
 import logging
 import os
-import re
-import sys
 import time
 import traceback
 from datetime import date, datetime
@@ -12,11 +10,11 @@
 import litellm
 from langfuse.decorators import langfuse_context, observe
 
-from config import get_settings
-from tools.search_hadith import SearchHadith
-from tools.search_vectara import SearchVectara
-from tools.search_quran import SearchQuran
-from util.prompt_mgr import PromptMgr
+from ansari.config import get_settings
+from ansari.tools.search_hadith import SearchHadith
+from ansari.tools.search_vectara import SearchVectara
+from ansari.tools.search_quran import SearchQuran
+from ansari.util.prompt_mgr import PromptMgr
 
 logger = logging.getLogger(__name__ + ".Ansari")
 logging_level = get_settings().LOGGING_LEVEL.upper()
@@ -49,7 +47,7 @@ def __init__(self, settings, message_logger=None, json_format=False):
             sm.get_tool_name(): sm,
         }
         self.model = settings.MODEL
-        self.pm = PromptMgr()
+        self.pm = PromptMgr(src_dir=settings.PROMPT_PATH)
         self.sys_msg = self.pm.bind(settings.SYSTEM_PROMPT_FILE_NAME).render()
         self.tools = [
             x.get_tool_description() for x in self.tool_name_to_instance.values()

agents/ansari_workflow.py → src/ansari/agents/ansari_workflow.py

@@ -6,10 +6,10 @@
 
 import litellm
 
-from tools.search_hadith import SearchHadith
-from tools.search_vectara import SearchVectara
-from tools.search_quran import SearchQuran
-from util.prompt_mgr import PromptMgr
+from ansari.tools.search_hadith import SearchHadith
+from ansari.tools.search_vectara import SearchVectara
+from ansari.tools.search_quran import SearchQuran
+from ansari.util.prompt_mgr import PromptMgr
 
 logger = logging.getLogger(__name__ + ".AnsariWorkflow")
 

ansari_db.py → src/ansari/ansari_db.py

@@ -2,15 +2,15 @@
 import logging
 from contextlib import contextmanager
 from datetime import datetime, timedelta, timezone
-
+from typing import Union
 import bcrypt
 import jwt
 import psycopg2
 import psycopg2.pool
 from fastapi import HTTPException, Request
 from jwt import ExpiredSignatureError, InvalidTokenError
 
-from config import Settings, get_settings
+from ansari.config import Settings, get_settings
 
 logger = logging.getLogger(__name__)
 logging_level = get_settings().LOGGING_LEVEL.upper()
@@ -588,7 +588,7 @@ def store_quran_answer(self, surah: int, ayah: int, question: str, ansari_answer
                 )
                 conn.commit()
 
-    def get_quran_answer(self, surah: int, ayah: int, question: str) -> str | None:
+    def get_quran_answer(self, surah: int, ayah: int, question: str) -> Union[str, None]:
         """
         Retrieve the stored answer for a given surah, ayah, and question.
 
@@ -612,7 +612,6 @@ def get_quran_answer(self, surah: int, ayah: int, question: str) -> str | None:
                     """
                     cur.execute(select_cmd, (surah, ayah, question))
                     result = cur.fetchone()
-
                     if result:
                         return result[0]
                     else:

main_api.py → src/ansari/app/main_api.py

@@ -1,5 +1,6 @@
 import logging
 import os
+from typing import Union
 import uuid
 
 import psycopg2
@@ -16,12 +17,12 @@
 from sendgrid.helpers.mail import Mail
 from zxcvbn import zxcvbn
 
-from agents.ansari import Ansari
-from agents.ansari_workflow import AnsariWorkflow
-from ansari_db import AnsariDB, MessageLogger
-from config import Settings, get_settings
-from main_whatsapp import router as whatsapp_router
-from presenters.api_presenter import ApiPresenter
+from ansari.agents import Ansari
+from ansari.agents import AnsariWorkflow
+from ansari.ansari_db import AnsariDB, MessageLogger
+from ansari.config import Settings, get_settings
+from ansari.app.main_whatsapp import router as whatsapp_router
+from ansari.presenters.api_presenter import ApiPresenter
 
 logger = logging.getLogger(__name__)
 logging_level = get_settings().LOGGING_LEVEL.upper()
@@ -36,7 +37,6 @@
 def main():
   add_app_middleware()
 
-
 def add_app_middleware():
     app.add_middleware(
         CORSMiddleware,
@@ -46,8 +46,7 @@ def add_app_middleware():
         allow_headers=["*"],
     )
 
-main() 
-
+main()
 db = AnsariDB(get_settings())
 ansari = Ansari(get_settings())
 
@@ -689,7 +688,7 @@ class AyahQuestionRequest(BaseModel):
     surah: int
     ayah: int
     question: str
-    augment_question: bool | None = False
+    augment_question: Union[bool,None] = False
     apikey: str
 
 @app.post("/api/v2/ayah")
@@ -703,6 +702,7 @@ async def answer_ayah_question(
 
     try:
         # Create AnsariWorkflow instance
+        logging.debug("Creating AnsariWorkflow instance for {req.surah}:{req.ayah}")
         ansari_workflow = AnsariWorkflow(settings)
 
         ayah_id = req.surah*1000 + req.ayah
@@ -750,6 +750,6 @@ async def answer_ayah_question(
         db.store_quran_answer(req.surah, req.ayah, req.question, ansari_answer)
 
         return {"response": ansari_answer}
-    except Exception as e:
-        logger.error(f"Error in answer_ayah_question: {str(e)}")
+    except Exception:
+        logger.error("Error in answer_ayah_question", exc_info=True)
         raise HTTPException(status_code=500, detail="Internal server error")

main_discord.py → src/ansari/app/main_discord.py

@@ -1,5 +1,5 @@
-from agents.ansari import Ansari
-from config import get_settings
+from ansari.agents import Ansari
+from ansari.config import get_settings
 from presenters.discord_presenter import DiscordPresenter
 
 # This work involves 3 agents, with Ansari as primary.

src/ansari/app/main_file.py

@@ -0,0 +1,9 @@
+import typer
+
+from ansari.agents import Ansari
+from ansari.config import get_settings
+from ansari.presenters.file_presenter import FilePresenter
+
+if __name__ == "__main__":
+    ansari = Ansari(get_settings())
+    typer.run(FilePresenter(ansari).present)

main_gradio.py → src/ansari/app/main_gradio.py

@@ -1,6 +1,6 @@
-from agents.ansari import Ansari
-from config import get_settings
-from presenters.gradio_presenter import GradioPresenter
+from ansari.agents import Ansari
+from ansari.config import get_settings
+from ansari.presenters.gradio_presenter import GradioPresenter
 
 if __name__ == "__main__":
     agent = Ansari(get_settings())

main_stdio.py → src/ansari/app/main_stdio.py

@@ -1,8 +1,8 @@
 import logging
 
-from agents.ansari import Ansari
-from config import get_settings
-from presenters.stdio_presenter import StdioPresenter
+from ansari.agents import Ansari
+from ansari.config import get_settings
+from ansari.presenters.stdio_presenter import StdioPresenter
 
 if __name__ == "__main__":
     logging.basicConfig(level=logging.INFO)

main_whatsapp.py → src/ansari/app/main_whatsapp.py

@@ -4,9 +4,9 @@
 from fastapi import APIRouter, HTTPException, Request
 from fastapi.responses import HTMLResponse
 
-from agents.ansari import Ansari
-from config import get_settings
-from presenters.whatsapp_presenter import WhatsAppPresenter
+from ansari.agents import Ansari
+from ansari.config import get_settings
+from ansari.presenters.whatsapp_presenter import WhatsAppPresenter
 
 # Initialize logging
 logger = logging.getLogger(__name__)

config.py → src/ansari/config.py

@@ -1,4 +1,5 @@
 import logging
+from pathlib import Path
 from functools import lru_cache
 from typing import Literal, Optional, Union
 
@@ -33,6 +34,15 @@ class Settings(BaseSettings):
         missing="ignore",
     )
 
+    def get_resource_path(filename):
+        # Get the directory of the current script
+        script_dir = Path(__file__).resolve()
+        # Construct the path to the resources directory
+        resources_dir = script_dir.parent / 'resources'
+        # Construct the full path to the resource file
+        path = resources_dir / filename
+        return path
+
     DATABASE_URL: PostgresDsn = Field(
         default="postgresql://postgres:password@localhost:5432/ansari"
     )
@@ -82,12 +92,12 @@ class Settings(BaseSettings):
     TAFSIR_FN_NAME: str = Field(default="search_tafsir")
     TAFSIR_FN_DESCRIPTION: str = Field(
         default="""
-        Queries Tafsir Ibn Kathir (the renowned Qur'anic exegesis) for relevant 
-        interpretations and explanations. You call this function when you need to 
-        provide authoritative Qur'anic commentary and understanding based on Ibn 
-        Kathir's work. Regardless of the language used in the original conversation, 
-        you will translate the query into English before searching the tafsir. The 
-        function returns a list of **potentially** relevant matches, which may include 
+        Queries Tafsir Ibn Kathir (the renowned Qur'anic exegesis) for relevant
+        interpretations and explanations. You call this function when you need to
+        provide authoritative Qur'anic commentary and understanding based on Ibn
+        Kathir's work. Regardless of the language used in the original conversation,
+        you will translate the query into English before searching the tafsir. The
+        function returns a list of **potentially** relevant matches, which may include
         multiple passages of interpretation and analysis.
         """
     )
@@ -114,13 +124,15 @@ class Settings(BaseSettings):
     WHATSAPP_ACCESS_TOKEN_FROM_SYS_USER: Optional[SecretStr] = Field(default=None)
     WHATSAPP_VERIFY_TOKEN_FOR_WEBHOOK: Optional[SecretStr] = Field(default=None)
 
-    template_dir: DirectoryPath = Field(default="resources/templates")
+    template_dir: DirectoryPath = Field(default=get_resource_path("templates"))
     diskcache_dir: str = Field(default="diskcache_dir")
 
     MODEL: str = Field(default="gpt-4o")
     MAX_TOOL_TRIES: int = Field(default=3)
     MAX_FAILURES: int = Field(default=1)
     SYSTEM_PROMPT_FILE_NAME: str = Field(default="system_msg_tool")
+    PROMPT_PATH: str = Field(default=str(get_resource_path("prompts")))
+
 
     LOGGING_LEVEL: str = Field(default="INFO")
 

presenters/whatsapp_presenter.py → src/ansari/presenters/whatsapp_presenter.py

@@ -6,7 +6,7 @@
 from fastapi import FastAPI
 from fastapi.middleware.cors import CORSMiddleware
 
-from config import get_settings
+from ansari.config import get_settings
 
 # Initialize logging
 logger = logging.getLogger(__name__)

tools/search_vectara.py → src/ansari/tools/search_vectara.py

@@ -1,4 +1,5 @@
 import json
+import logging
 import requests
 
 

src/ansari/util/__init__.py

@@ -0,0 +1,4 @@
+# This file makes the 'util' directory a package.
+from .prompt_mgr import PromptMgr
+
+__all__ = ["PromptMgr"]

util/prompt_mgr.py → src/ansari/util/prompt_mgr.py

@@ -1,6 +1,7 @@
 from typing import Union
 
 from pydantic import BaseModel
+from pathlib import Path
 
 
 class Prompt(BaseModel):
@@ -16,7 +17,16 @@ def render(self, **kwargs) -> str:
 
 
 class PromptMgr:
-    def __init__(self, hot_reload: bool = True, src_dir: str = "resources/prompts"):
+    def get_resource_path(filename):
+        # Get the directory of the current script
+        script_dir = Path(__file__).resolve()
+        # Construct the path to the resources directory
+        resources_dir = script_dir.parent.parent / 'resources'
+        # Construct the full path to the resource file
+        path = resources_dir / filename
+        return path
+
+    def __init__(self, hot_reload: bool = True, src_dir: str = str(get_resource_path("prompts"))):
         """Creates a prompt manager.
 
         Args: