user2595/ci-ai
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

chore(llm): refresh defaults to match deployed Jetson backend #2

Merged
user2595 merged 2 commits from chore/llm-defaults-jetson into main 2026-06-14 06:22:05 +00:00
Conversation 0 Commits 2 Files Changed 6 +70 -20
user2595 commented 2026-06-13 13:19:02 +00:00
Owner
Copy Link

Summary

  • scripts/llm.py defaulted to AI_PROVIDER=anthropic, AI_MODEL=qwen-light (ollama), and the dead .20 Jetson IP — none of which match how the CI workflows and ci-ai-bot actually run it.
  • Update defaults to the live self-hosted setup: AI_PROVIDER=ollama, AI_MODEL=coder (qwen2.5-coder:7b), AI_BASE_URL=http://192.168.178.103:11434/v1.
  • Refresh the module docstring (Jetson is now the default backend, anthropic the optional alternative; drop the stale "jetzt/später" framing and .20 reference).
  • Update the README config table/note that flagged this exact staleness ("defaults refresh pending") now that defaults and runtime agree.

Verification

  • python3 -m py_compile scripts/llm.py passes.
  • Grepped .gitea/workflows/, workflows/*.example.yml, bot/, and ansible-eulernest's docker-compose.yml.j2 / group_vars/pi/vars.yml — all set AI_PROVIDER/AI_MODEL/AI_BASE_URL explicitly via env, so this default change is purely cosmetic for them and only affects runs with no env vars set.

Test plan

  • CI / AI review workflow still runs (env vars unchanged, so behavior should be identical)
## Summary - `scripts/llm.py` defaulted to `AI_PROVIDER=anthropic`, `AI_MODEL=qwen-light` (ollama), and the dead `.20` Jetson IP — none of which match how the CI workflows and ci-ai-bot actually run it. - Update defaults to the live self-hosted setup: `AI_PROVIDER=ollama`, `AI_MODEL=coder` (qwen2.5-coder:7b), `AI_BASE_URL=http://192.168.178.103:11434/v1`. - Refresh the module docstring (Jetson is now the default backend, anthropic the optional alternative; drop the stale "jetzt/später" framing and `.20` reference). - Update the README config table/note that flagged this exact staleness ("defaults refresh pending") now that defaults and runtime agree. ## Verification - `python3 -m py_compile scripts/llm.py` passes. - Grepped `.gitea/workflows/`, `workflows/*.example.yml`, `bot/`, and ansible-eulernest's `docker-compose.yml.j2` / `group_vars/pi/vars.yml` — all set `AI_PROVIDER`/`AI_MODEL`/`AI_BASE_URL` explicitly via env, so this default change is purely cosmetic for them and only affects runs with no env vars set. ## Test plan - [ ] CI / AI review workflow still runs (env vars unchanged, so behavior should be identical)
user2595 added 2 commits 2026-06-13 13:19:03 +00:00
docs: document interactive ci-ai-bot + current Jetson backend 1c84d7fc2a 
- README: new Feature F (interactive @ci-ai webhook bot) — table, tree, dedicated
  section (events, build, deploy pointer, env); reflect the Jetson (Ollama, coder,
  .103) as the live backend instead of "Claude now / Jetson later"
- flag that llm.py code defaults (anthropic / .20 / qwen-light) are stale vs. the
  configured ollama/coder/.103 — defaults refresh pending
- new bot/README.md

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
chore(llm): refresh defaults to match deployed Jetson backend 688699a6ad 
llm.py defaulted to AI_PROVIDER=anthropic, AI_MODEL=qwen-light, and the
dead .20 Jetson IP — none of which match how workflows/bot actually run
it (ollama / coder / .103). Update the defaults and docstring to reflect
the live self-hosted setup, and refresh the README note that flagged the
mismatch.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
user2595 added 2 commits 2026-06-13 13:20:27 +00:00
chore(llm): refresh defaults to match deployed Jetson backend 688699a6ad 
llm.py defaulted to AI_PROVIDER=anthropic, AI_MODEL=qwen-light, and the
dead .20 Jetson IP — none of which match how workflows/bot actually run
it (ollama / coder / .103). Update the defaults and docstring to reflect
the live self-hosted setup, and refresh the README note that flagged the
mismatch.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
chore(llm): refresh defaults to match deployed Jetson backend 72a902010a 
llm.py defaulted to AI_PROVIDER=anthropic, AI_MODEL=qwen-light, and the
dead .20 Jetson IP — none of which match how workflows/bot actually run
it (ollama / coder / .103). Update the defaults and docstring to reflect
the live self-hosted setup, and refresh the README note that flagged the
mismatch.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
user2595 added 1 commit 2026-06-14 06:19:09 +00:00
feat(cd): auto-deploy ci-ai-bot on push + dehard-code Jetson URL 38a6851b46 
- .gitea/workflows/cd.yml: on push to main (bot/** or scripts/**),
  build ci-ai-bot:latest on the Pi runner via DooD and force-recreate
  the container with docker compose (no registry push needed)
- reusable-{ai-review,commands,security}.yml: replace hardcoded
  192.168.178.103:11434 with ${{ vars.JETSON_OLLAMA_URL }} (org variable
  set in conformallab → fallback to old IP if variable is missing)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
user2595 merged commit 6ff95287fa into main 2026-06-14 06:22:05 +00:00
user2595 deleted branch chore/llm-defaults-jetson 2026-06-14 06:22:05 +00:00
Sign in to join this conversation.
Reviewers
No Reviewers
Labels
No items
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
user2595
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: user2595/ci-ai#2
No description provided.
Delete Branch "chore/llm-defaults-jetson"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?