# Copy to .env and fill in real values. .env is gitignored.
#
# The LLM clustering pipeline (templates/pipeline/cluster_themes.py) uses an
# OpenAI-compatible API. Any provider that speaks /chat/completions works:
#
#   DeepSeek   OPENAI_BASE_URL=https://api.deepseek.com/v1          OPENAI_MODEL=deepseek-chat
#   Qwen       OPENAI_BASE_URL=https://dashscope.aliyuncs.com/compatible-mode/v1  OPENAI_MODEL=qwen-plus
#   Kimi       OPENAI_BASE_URL=https://api.moonshot.cn/v1           OPENAI_MODEL=moonshot-v1-8k
#   GLM        OPENAI_BASE_URL=https://open.bigmodel.cn/api/paas/v4 OPENAI_MODEL=glm-4
#   Ollama     OPENAI_BASE_URL=http://localhost:11434/v1            OPENAI_MODEL=qwen2.5:14b
#   OpenAI     OPENAI_BASE_URL=https://api.openai.com/v1            OPENAI_MODEL=gpt-4o-mini

OPENAI_BASE_URL=https://api.deepseek.com/v1
OPENAI_API_KEY=sk-your-key-here
OPENAI_MODEL=deepseek-chat

# Reddit official API (optional but STRONGLY recommended).
# Without these, the collector scrapes www.reddit.com unauthenticated — which
# is frequently IP-blocked (403) and returns nothing, starving consumer-product
# reports of their main data source. With them, the collector uses the official
# oauth.reddit.com API: works even when the web frontend blocks your IP, much
# higher rate limits, far more voices per report.
# Setup (~5 min): see docs/reddit-oauth.md
REDDIT_CLIENT_ID=
REDDIT_CLIENT_SECRET=
# Optional but most reliable for "script" apps: your Reddit login.
# If set, uses password grant (Reddit's official path for script apps).
# Leave blank to use app-only client_credentials (id+secret only).
REDDIT_USERNAME=
REDDIT_PASSWORD=