Files
ERP/bot/app/config.py
T
whqreza d58b619107 feat(observability): phase-1 silent pipeline recorder + phase-0 Jev validation
Phase 0 — Jev validated with the real key (5 Persian scenarios, api.typesafe.ai): stale project name «منشیانه»->«منشانه» conf 0.73 (ambiguity band), STT spelling variants «مصطفوی/مستفدی/مسطفی»->«بازرگانی مصطفوی» conf 0.93, exact «دفتر» 0.88, proper _none rejection for «کاغذ A4» 0.92, and two mentions in ONE request (0.77/0.64). Exact model version to pin: jev-1.13.0. Latency 0.7-1.1s per field request; ~450-650 in / 64-133 out tokens.

Phase 1 — Observability (DB-only, zero extra LLM tokens, nothing is ever deleted):

- new tables bot_traces + bot_trace_events (idempotent DDL in db_init.py; ORM in models.py)

- new app/observability.py: _jsonable (UUID/datetime/Decimal/set/enum/dataclass-safe), trace_start, event, log_resolution, link_draft, latest_trace_for_draft, set_trace_status, span() context manager, save_voice (Path.mkdir parents+exist_ok, day folders)

- capture points wired: input.text; input.voice (+voice persisted under OBS_VOICE_DIR); stt.run + per-provider stt.result (success AND failure, latency, model); nlu.request + nlu.result with prompt_hash/prompt_version; entities.extracted; routing.decision (difflib baseline for the phase-6 A/B); ambiguity.sent; ambiguity.user_choice (buttons + free-text reply path); user_choice.pick; draft.render; draft.edit (wizard-side edits detected by the sync loop); draft.commit (quote/supplier/material/project + trace status/quote id); draft.cancel

- compose: persistent ./bot/data:/srv/bot/data volume; bot/.gitignore: data/ (runtime data never committed, never deleted)

- tests: bot/tools/test_observability.py -> 8/8 pass (jsonable cases, trace lifecycle, JSONB round-trip, draft linking, status update, save_voice day-folder mkdir, span task). Verified live: voice file lands on the host volume and is gitignored; rows + files persist across bot restart.
2026-10-12 21:04:50 +03:30

155 lines
5.4 KiB
Python

"""Bot configuration.
All settings can be overridden with environment variables prefixed with
``PROCUREMENT_BOT_`` (e.g. ``PROCUREMENT_BOT_BOT_TOKEN=...``) or a local
``.env`` file. Docker/compose passes them in directly (same convention as
the api service's ``PROCUREMENT_*`` prefix).
"""
from pydantic_settings import BaseSettings, SettingsConfigDict
class Settings(BaseSettings):
model_config = SettingsConfigDict(
env_prefix="PROCUREMENT_BOT_",
env_file=".env",
env_file_encoding="utf-8",
extra="ignore",
)
# Bale Bot API — Telegram-compatible clone hosted at tapi.bale.ai.
BALE_API_BASE: str = "https://tapi.bale.ai"
BOT_TOKEN: str = "" # from Bale's BotFather
POLL_TIMEOUT_SECONDS: int = 25 # getUpdates long-poll window
POLL_ERROR_SLEEP_SECONDS: float = 3.0
# ERP API (in-network service URL inside Docker).
API_BASE_URL: str = "http://localhost:8000/api/v1"
API_TIMEOUT_SECONDS: float = 30.0
# Optional bearer for the day procurement endpoints start enforcing auth.
SERVICE_TOKEN: str = ""
# Frontend URL (for opening Wizard from Bale bot)
FRONTEND_URL: str = "http://localhost:8081"
# Security token for standalone wizard review URL
DRAFT_TOKEN_SECRET: str = "procurement-draft-secret-change-in-prod-v7"
DRAFT_TOKEN_EXPIRE_SECONDS: int = 7200 # 2 hours
# Persistent storage directory for draft attachments (proformas)
ATTACHMENTS_DIR: str = "data/attachments/proformas"
# OpenAI-compatible AI provider (LLM), e.g. OpenAI / GLM / DeepSeek / OpenRouter.
AI_BASE_URL: str = "" # e.g. https://api.openai.com/v1
AI_API_KEY: str = ""
LLM_MODEL: str = "gpt-4o-mini"
LLM_TIMEOUT_SECONDS: float = 60.0
# STT providers — support multiple models/APIs in parallel.
# 1. JSON list: [{"name": "m1", "base_url": "...", "api_key": "...", "model": "..."}, ...]
STT_PROVIDERS_JSON: str = ""
# 2. Individual provider variables (up to 3 distinct STT models)
STT_PROVIDER_1_NAME: str = "model-1"
STT_PROVIDER_1_BASE_URL: str = ""
STT_PROVIDER_1_API_KEY: str = ""
STT_PROVIDER_1_MODEL: str = ""
STT_PROVIDER_1_LANGUAGE: str = "fa"
STT_PROVIDER_2_NAME: str = "model-2"
STT_PROVIDER_2_BASE_URL: str = ""
STT_PROVIDER_2_API_KEY: str = ""
STT_PROVIDER_2_MODEL: str = ""
STT_PROVIDER_2_LANGUAGE: str = "fa"
STT_PROVIDER_3_NAME: str = "model-3"
STT_PROVIDER_3_BASE_URL: str = ""
STT_PROVIDER_3_API_KEY: str = ""
STT_PROVIDER_3_MODEL: str = ""
STT_PROVIDER_3_LANGUAGE: str = "fa"
# Legacy single STT provider fallback:
STT_BASE_URL: str = ""
STT_API_KEY: str = ""
STT_MODEL: str = "whisper-1"
STT_LANGUAGE: str = "fa"
@property
def stt_base_url(self) -> str:
return self.STT_BASE_URL or self.AI_BASE_URL
@property
def stt_api_key(self) -> str:
return self.STT_API_KEY or self.AI_API_KEY
@property
def stt_providers(self) -> list[dict]:
import json
if self.STT_PROVIDERS_JSON:
try:
data = json.loads(self.STT_PROVIDERS_JSON)
if isinstance(data, list) and len(data) > 0:
return data
except Exception:
pass
# Check explicit STT_PROVIDER_1 / 2 / 3
custom_list: list[dict] = []
for idx in (1, 2, 3):
b_url = getattr(self, f"STT_PROVIDER_{idx}_BASE_URL")
key = getattr(self, f"STT_PROVIDER_{idx}_API_KEY")
model = getattr(self, f"STT_PROVIDER_{idx}_MODEL")
name = getattr(self, f"STT_PROVIDER_{idx}_NAME")
lang = getattr(self, f"STT_PROVIDER_{idx}_LANGUAGE") or "fa"
if b_url and key:
custom_list.append({
"name": name or f"model-{idx}",
"base_url": b_url,
"api_key": key,
"model": model or self.STT_MODEL,
"language": lang,
"timeout": 30.0,
})
if custom_list:
return custom_list
# Fallback to single legacy STT provider if set
if self.stt_base_url and self.stt_api_key:
return [{
"name": "default",
"base_url": self.stt_base_url,
"api_key": self.stt_api_key,
"model": self.STT_MODEL,
"language": self.STT_LANGUAGE,
"timeout": 30.0,
}]
return []
# PostgreSQL — same instance/database as the api service.
DATABASE_URL: str = "postgresql+asyncpg://postgres:postgres@localhost:5432/procurement"
# Chats that may use the bot without Casdoor linking (comma separated).
ADMIN_CHAT_IDS: str = ""
# Drafts older than this are marked expired by the sweeper loop.
DRAFT_TTL_HOURS: int = 24
# Observability (phase 1) — silent pipeline recorder (DB only; no external
# tools, no extra LLM tokens). Data is NEVER deleted.
OBS_ENABLED: bool = True
OBS_VOICE_DIR: str = "data/voices" # persistent volume (./bot/data) — keep forever
OBS_PROMPT_VERSION: str = "" # manual label for the NLU prompt in use
LOG_LEVEL: str = "INFO"
@property
def admin_chat_ids(self) -> set[int]:
out: set[int] = set()
for part in self.ADMIN_CHAT_IDS.replace(";", ",").split(","):
part = part.strip()
if part.isdigit():
out.add(int(part))
return out
settings = Settings()