rene/tip-training-data
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

crawl: add robot-status training records [2026-04-29T20:11:24.091Z]

Browse Source
This commit is contained in:
TIP Crawler 2026-04-29 22:11:24 +02:00
parent b8ec33a09b
commit f1c83f8983
2 changed files with 2 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
qa-pairs/robot-control-high.jsonl Normal file
View File

@ -0,0 +1 @@
{"id":"robot-exp-f3015c7c20d8","source":"robot-experience:status:codex-verification-robot","kind":"sft-jsonl","messages":[{"role":"system","content":"You are TIPLLM, an expert AI assistant for the Transceiver Intelligence Platform (TIP). You have deep knowledge of optical transceiver specifications, form factors, IEEE standards, vendor product catalogs, and fiber optic networking. You help engineers select, source, and validate transceivers. You provide precise, structured answers with confidence scores and always cite your reasoning."},{"role":"user","content":"You are TIPLLM controlling TIP crawler and verification robots.\nGiven this current observation, choose the safest next robot action.\n\n{\n \"event\": \"status\",\n \"actor\": \"codex-verification-robot\",\n \"input\": {\n \"live_total\": 13546,\n \"price_verified\": 7250,\n \"image_verified\": 7025,\n \"details_verified\": 6243,\n \"fully_verified\": 5812,\n \"requested_policy\": \"TIPLLM only; no other AI; record experiences in Gitea training pool\"\n },\n \"safety_notes\": [\n \"No crawler jobs were started.\",\n \"No non-TIPLLM AI is added to the runtime design.\",\n \"This record is intended for TIPLLM SFT training.\"\n ]\n}"},{"role":"assistant","content":"{\n \"summary\": \"TIP verification robot policy updated: use TIPLLM only for crawler planning and write robot experience back to the TIPLLM training pool.\",\n \"decision\": {\n \"ai_policy\": \"TIPLLM_ONLY\",\n \"erik_policy\": \"controller_only_no_heavy_crawlers\",\n \"training_pool\": \"write robot status, TIPLLM plans, queue plans, enqueue outcomes and crawler results as SFT records\",\n \"deployment_policy\": \"Proxmox/Pi for heavy crawler work; Erik only safe queues and DB/pg-boss coordination\"\n },\n \"outcome\": {\n \"code_paths\": [\n \"packages/scraper/src/crawler-llm/training-data-writer.ts\",\n \"packages/scraper/src/robots/verification-robots.ts\",\n \"scripts/tip-learning-pool-build.ts\"\n ],\n \"crawl_jobs_started\": 0\n },\n \"constraints\": [\n \"Use TIPLLM only for planning/extraction feedback.\",\n \"Do not start heavy crawler waves on Erik.\",\n \"Prefer small, observable queue batches.\",\n \"Write useful outcomes back into the TIPLLM training pool.\"\n ]\n}"}]}

1
robot-experiences/2026-04-29.jsonl Normal file
View File

@ -0,0 +1 @@
{"event":"status","observed_at":"2026-04-29T20:11:24.073Z","actor":"codex-verification-robot","summary":"TIP verification robot policy updated: use TIPLLM only for crawler planning and write robot experience back to the TIPLLM training pool.","input":{"live_total":13546,"price_verified":7250,"image_verified":7025,"details_verified":6243,"fully_verified":5812,"requested_policy":"TIPLLM only; no other AI; record experiences in Gitea training pool"},"decision":{"ai_policy":"TIPLLM_ONLY","erik_policy":"controller_only_no_heavy_crawlers","training_pool":"write robot status, TIPLLM plans, queue plans, enqueue outcomes and crawler results as SFT records","deployment_policy":"Proxmox/Pi for heavy crawler work; Erik only safe queues and DB/pg-boss coordination"},"outcome":{"code_paths":["packages/scraper/src/crawler-llm/training-data-writer.ts","packages/scraper/src/robots/verification-robots.ts","scripts/tip-learning-pool-build.ts"],"crawl_jobs_started":0},"safety_notes":["No crawler jobs were started.","No non-TIPLLM AI is added to the runtime design.","This record is intended for TIPLLM SFT training."]}