File size: 2,249 Bytes
519b145
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
#############################################
## Django / Local Development
#############################################
DJANGO_SECRET_KEY=change-me-in-development
DJANGO_DEBUG=true
DJANGO_ALLOWED_HOSTS=localhost,127.0.0.1

#############################################
## Local PostgreSQL (Docker compose defaults)
#############################################
POSTGRES_HOST=localhost
POSTGRES_PORT=5543
POSTGRES_DB=hue_portal
POSTGRES_USER=hue
POSTGRES_PASSWORD=huepass

#############################################
## Redis Cache (Optional - for query rewrite and prefetch caching)
#############################################
# Supports Upstash and Railway Redis free tier
REDIS_URL=redis://localhost:6380/0
# Cache TTLs (in seconds)
CACHE_QUERY_REWRITE_TTL=3600  # 1 hour
CACHE_PREFETCH_TTL=1800  # 30 minutes

#############################################
## Hugging Face / Tunnel automation
#############################################
HF_SPACE_ID=davidtran999/hue-portal-backend
# Nếu không export HF_TOKEN trong shell, tool sẽ cố đọc ~/.cache/huggingface/token
HF_TOKEN=

# Ngrok / Cloudflare tunnel settings
NGROK_BIN=ngrok
NGROK_REGION=ap
NGROK_AUTHTOKEN=
PG_TUNNEL_LOCAL_PORT=5543
PG_TUNNEL_WATCH_INTERVAL=45

# Credentials that sẽ được đẩy lên HF secrets
PG_TUNNEL_USER=hue_remote
PG_TUNNEL_PASSWORD=huepass123
PG_TUNNEL_DB=hue_portal

#############################################
## LLM / llama.cpp (Qwen2.5-1.5b or Vi-Qwen2-3B-RAG) defaults
#############################################
DEFAULT_LLM_PROVIDER=llama_cpp
LLM_PROVIDER=llama_cpp
# Model path (local file path or Hugging Face repo)
LLM_MODEL_PATH=/app/backend/models/qwen2.5-1.5b-instruct-q5_k_m.gguf
# Future: Vi-Qwen2-3B-RAG (when Phase 3 is complete)
# LLM_MODEL_PATH=/app/backend/models/vi-qwen2-3b-rag-q5_k_m.gguf
LLAMA_CPP_CONTEXT=4096
LLAMA_CPP_THREADS=2
LLAMA_CPP_BATCH=512
LLAMA_CPP_MAX_TOKENS=512
LLAMA_CPP_TEMPERATURE=0.35
LLAMA_CPP_TOP_P=0.85
LLAMA_CPP_REPEAT_PENALTY=1.1
LLAMA_CPP_USE_MMAP=true
LLAMA_CPP_USE_MLOCK=true
RUN_HEAVY_STARTUP_TASKS=0

#############################################
## Frontend
#############################################
# Gán VITE_API_BASE khi muốn trỏ tới API khác (vd HF Space)
VITE_API_BASE=