Explorar o código

.env 入库(内网团队约定)+ 补 5 个 CK_* 新键;gitignore 不再忽略 .env(pem 仍忽略)

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
SamLee hai 3 semanas
pai
achega
36264c60fa
Modificáronse 2 ficheiros con 209 adicións e 3 borrados
  1. 208 0
      .env
  2. 1 3
      .gitignore

+ 208 - 0
.env

@@ -0,0 +1,208 @@
+# ContentFindAgentNew environment variables
+# Generated from tech_documents/data_interface.md and legacy ContentFindAgent code anchors.
+# Do not commit real secrets. Fill blank values from the approved secrets store before running live jobs.
+
+# -----------------------------------------------------------------------------
+# Runtime
+# -----------------------------------------------------------------------------
+HOST=0.0.0.0
+PORT=8080
+TZ=Asia/Shanghai
+SCHEDULER_TIMEZONE=Asia/Shanghai
+MAX_CONCURRENT_TASKS=1
+SCHEDULE_DISPATCH_INTERVAL_SECONDS=30
+SCHEDULE_TASK_TIMEOUT_SECONDS=1500
+
+TRACE_DIR=.cache/traces
+OUTPUT_DIR=.cache/output
+INPUT_LOG_PATH=.cache/input_log
+OUTPUT_HTML_PATH=
+OUTPUT_JSON_PATH=
+TRACE_ID=
+SKILLS_DIR=./skills
+ENABLED_SKILLS=
+
+MODEL=anthropic/claude-sonnet-4.5
+TEMPERATURE=0.3
+MAX_ITERATIONS=30
+AGENT_PRICING_CONFIG=
+AGENT_DISABLE_SIDE_BRANCHES=
+AGENT_DEBUG=
+
+# -----------------------------------------------------------------------------
+# LLM providers
+# -----------------------------------------------------------------------------
+OPEN_ROUTER_API_KEY=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
+OPENROUTER_API_KEY=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
+OPENROUTER_BASE_URL=https://openrouter.ai/api/v1
+# 内容判定:Gemini 直读视频(V3-M2),复用上面 OPENROUTER_API_KEY/BASE_URL
+CONTENT_AGENT_VIDEO_LLM_MODEL=google/gemini-3-flash-preview
+CONTENT_AGENT_VIDEO_LLM_TIMEOUT_SECONDS=90
+
+QWEN_API_KEY=
+QWEN_BASE_URL=https://dashscope.aliyuncs.com/compatible-mode/v1
+
+GEMINI_API_KEY=
+YESCODE_BASE_URL=
+YESCODE_API_KEY=
+
+# -----------------------------------------------------------------------------
+# Upstream open_aigc / Pattern V2 inventory
+# Used by open_aigc.post, post_decode_topic_point_element,
+# post_script_paragraph_field_element, global_category, global_element,
+# element_classification_mapping, pattern_mining_execution,
+# pattern_mining_element, pattern_itemset, pattern_itemset_item.
+# Fill only when v2 service actually connects to these upstream stores.
+# -----------------------------------------------------------------------------
+OPEN_AIGC_PG_HOST=gp-t4n72471pkmt4b9q7o-master.gpdbmaster.singapore.rds.aliyuncs.com
+OPEN_AIGC_PG_PORT=5432
+OPEN_AIGC_PG_USER=aiddit_aigc
+OPEN_AIGC_PG_PASSWORD=%a&&yqNxg^V1$toJ*WOa^-b^X=QJ
+OPEN_AIGC_PG_DB_NAME=open_aigc
+OPEN_AIGC_PG_SCHEMA=public
+
+# -----------------------------------------------------------------------------
+# DemandAgent demand_content / CFA production fact DB
+# V1 reads DemandAgent demand_content and writes content_agent_* tables here.
+# Host uses the VPC private IP when CFA runs from this repo / local VPN.
+# -----------------------------------------------------------------------------
+CONTENT_SUPPLY_DB_HOST=192.168.82.27
+CONTENT_SUPPLY_DB_PORT=3306
+CONTENT_SUPPLY_DB_NAME=content-deconstruction-supply
+CONTENT_SUPPLY_DB_USER=content_rw
+CONTENT_SUPPLY_DB_PASSWORD=bC1aH4bA1lB0
+CONTENT_AGENT_DB_RUNTIME_ENABLED=0
+
+# -----------------------------------------------------------------------------
+# Platform APIs: Crawapi / Douyin(V3 双渠道 抖音+视频号 均走此 host)
+# -----------------------------------------------------------------------------
+CONTENTFIND_API_CRAWAPI_BASE_URL=http://crawler.aiddit.com
+CONTENTFIND_API_CRAWAPI_KEY=
+CONTENTFIND_API_CRAWAPI_TIMEOUT_SECONDS=60
+
+CONTENTFIND_DOUYIN_KEYWORD_PATH=/crawler/dou_yin/keyword
+CONTENTFIND_DOUYIN_BLOGGER_PATH=/crawler/dou_yin/blogger
+CONTENTFIND_DOUYIN_DETAIL_PATH=/crawler/dou_yin/detail
+
+CONTENTFIND_DOUYIN_DEFAULT_ACCOUNT_ID=771431222
+CONTENTFIND_DOUYIN_DEFAULT_CONTENT_TYPE=视频
+CONTENTFIND_DOUYIN_DEFAULT_SORT_TYPE=综合排序
+CONTENTFIND_DOUYIN_DEFAULT_PUBLISH_TIME=不限
+CONTENTFIND_DOUYIN_DEFAULT_CURSOR=0
+CONTENTFIND_DOUYIN_MAX_RESULTS_PER_QUERY=3
+CONTENTFIND_DOUYIN_ACCOUNT_WORKS_DEFAULT_SORT_TYPE=最新
+
+# -----------------------------------------------------------------------------
+# Good case / Pattern query tuning
+# -----------------------------------------------------------------------------
+GET_GOODCASE_TOPIC_POINT_IN_BATCH_SIZE=200
+GET_VIDEO_TOPIC_IN_BATCH_SIZE=200
+
+# -----------------------------------------------------------------------------
+# Scheduler external query source
+# -----------------------------------------------------------------------------
+SCHEDULE_QUERY_API=
+SCHEDULE_QUERY_API_KEY=
+SCHEDULE_QUERY_API_TIMEOUT=10.0
+
+# -----------------------------------------------------------------------------
+# AIGC plan handoff
+# Current evidence: source-located / blocked because this has write-side effects.
+# Keep CAN_NOT_CREATE_PLAN=true unless an operator explicitly enables live writes.
+# -----------------------------------------------------------------------------
+CAN_NOT_CREATE_PLAN=true
+CONTENTFIND_API_AIGC_BASE_URL=https://aigc-api.aiddit.com
+CONTENTFIND_API_AIGC_TOKEN=8bf14f27fc3a486788f3383452422d72
+AIGC_TOKEN=8bf14f27fc3a486788f3383452422d72
+PRODUCE_PLAN_ID=20260408085649635441036
+
+CONTENTFIND_AIGC_CRAWLER_PLAN_SAVE_PATH=/aigc/crawler/plan/save
+CONTENTFIND_AIGC_PRODUCE_PLAN_DETAIL_PATH=/aigc/produce/plan/detail
+CONTENTFIND_AIGC_PRODUCE_PLAN_SAVE_PATH=/aigc/produce/plan/save
+
+# -----------------------------------------------------------------------------
+# OSS log HTML upload
+# Current evidence: source-located / blocked because this has upload side effects.
+# -----------------------------------------------------------------------------
+ALIYUN_OSS_ACCESS_KEY_ID=
+ALIYUN_OSS_ACCESS_KEY_SECRET=
+ALIYUN_OSS_BUCKET=
+ALIYUN_OSS_REGION=
+ALIYUN_OSS_PREFIX=
+ALIYUN_OSS_PUBLIC_BASE_URL=
+
+# -----------------------------------------------------------------------------
+# Strategy learning external performance stores
+# Used by loghubods.video_dimension_detail_add_column and other performance review data.
+# Fill only when the strategy-learning layer connects to ODPS/Hive.
+# -----------------------------------------------------------------------------
+ODPS_ACCESS_ID=LTAI9EBa0bd5PrDa
+ODPS_ACCESS_KEY=vAalxds7YxhfOA2yVv8GziCg3Y87v5
+ODPS_ENDPOINT=http://service.odps.aliyun.com/api
+ODPS_PROJECT=loghubods
+ODPS_TABLE_VIDEO_DIMENSION_DETAIL=loghubods.video_dimension_detail_add_column
+
+HIVE_HOST=
+HIVE_PORT=
+HIVE_USER=
+HIVE_PASSWORD=
+HIVE_DATABASE=
+
+# -----------------------------------------------------------------------------
+# Optional framework services
+# -----------------------------------------------------------------------------
+KNOWHUB_API=http://localhost:8000
+BRAND_NAME=KnowHub
+BRAND_API_ENV=KNOWHUB_API
+BRAND_DB=knowhub.db
+ORG_KEYS=
+BROWSER_USE_API_KEY=
+
+# -----------------------------------------------------------------------------
+# Optional Feishu tools in the shared agent framework
+# Leave blank unless this project actually enables Feishu tools.
+# -----------------------------------------------------------------------------
+FEISHU_APP_ID=cli_a90fe317987a9cc9
+FEISHU_APP_SECRET=nn2dWuXTiRA2N6xodbm4g0qz1AfM2ayi
+FEISHU_AGENT_MODEL=qwen3.5-397b-a17b
+FEISHU_AGENT_SYSTEM_PROMPT=
+
+# -----------------------------------------------------------------------------
+# Vite / show frontend
+# Current show app is a static product sandbox. These are reserved for future API wiring.
+# -----------------------------------------------------------------------------
+VITE_API_BASE_URL=
+VITE_CONTENTFIND_API_BASE_URL=
+
+# -----------------------------------------------------------------------------
+# Added aliases from legacy implementation and data-engineering verification
+# -----------------------------------------------------------------------------
+CONTENTFIND_AIGC_READONLY_PRODUCE_PLAN_ID=20260408085649635441036
+CONTENTFIND_API_READONLY_TOKEN=8bf14f27fc3a486788f3383452422d72
+DATA_ENGINEERING_OPEN_AIGC_PG_DSN=postgresql://aiddit_aigc:%25a%26%26yqNxg%5EV1%24toJ%2AWOa%5E-b%5EX%3DQJ@gp-t4n72471pkmt4b9q7o-master.gpdbmaster.singapore.rds.aliyuncs.com:5432/open_aigc
+PGVECTOR_DSN=postgresql://aiddit_aigc:%25a%26%26yqNxg%5EV1%24toJ%2AWOa%5E-b%5EX%3DQJ@gp-t4n72471pkmt4b9q7o-master.gpdbmaster.singapore.rds.aliyuncs.com:5432/open_aigc
+OSS_ACCESS_KEY_ID=
+OSS_ACCESS_KEY_SECRET=
+OPEN_AIGC_PG_DATABASE=open_aigc
+
+# 火山方舟 embedding (Doubao-embedding-vision, 实测 2026-06-17)
+ARK_API_KEY=ark-6ca740e2-a719-4478-a8f8-bbb97985a784-059ce
+ARK_EMBEDDING_EP=ep-20260617180207-vmwg6
+ARK_EMBEDDING_URL=https://ark.cn-beijing.volces.com/api/v3/embeddings/multimodal
+ARK_EMBEDDING_DIM=2048
+
+# ── query 搜索 + OSS 转存(query→搜索→OSS→模型)实测 2026-06-26 ──
+# 关键词搜索:复用 crawler host(CONTENTFIND_API_CRAWAPI_BASE_URL=http://crawler.aiddit.com)
+#   小红书 POST /crawler/xiao_hong_shu/keyword  body{keyword,content_type:图文|视频,sort_type:综合,publish_time,cursor}
+#     → {code:0,data:{has_more,next_cursor,data:[{id:<content_id>,note_card:{...}}]}}
+#   抖音   POST /crawler/dou_yin/keyword       body{keyword,content_type:视频|图文,sort_type:最多点赞,cursor}
+#     ⚠ 实测 code 10000 未知错误,疑似需 cookie/鉴权或不稳定,待核实
+#   快手   POST /crawler/kuai_shou/keyword     body{keyword}
+# OSS 转存:把帖子媒体直链转存到公网 CDN,喂模型时用 cdn_url(绕过小红书防盗链 / 抖音临时签名链时效)
+#   POST <CK_OSS_UPLOAD_URL>  body{src_url,src_type:video|image}
+#     → {status:0,oss_object:{cdn_url:https://res.cybertogether.net/...,oss_object_key,save_oss_timestamp}}
+CK_OSS_UPLOAD_URL=http://crawler-upload-v2.aiddit.com/crawler/oss/upload_stream
+CK_OSS_UPLOAD_TIMEOUT_SECONDS=60
+CK_SEARCH_DEFAULT_LIMIT=5
+CK_SEARCH_CONTENT_TYPE=图文
+CK_SEARCH_SORT_TYPE=综合

+ 1 - 3
.gitignore

@@ -1,6 +1,4 @@
-# 密钥/环境 —— 绝不入库
-.env
-.env.*
+# 密钥 —— pem 绝不入库;.env 内网团队约定可入库
 *.pem
 
 # Python