fix(models): strip :cloud/-cloud suffix from models.dev Ollama Cloud IDs

[briandevans]

Summary

• fetch_ollama_cloud_models() merges live API results with models.dev additions
• models.dev appends :cloud / -cloud suffixes to Ollama Cloud IDs (e.g. kimi-k2.6:cloud, qwen3-coder:480b-cloud) that the live API does not use
• Without normalisation, these suffixed IDs bypass the dedup check and appear alongside the correct clean IDs in /model and hermes model

The bug

models.dev lists Ollama Cloud model IDs with appended :cloud / -cloud suffixes. The live Ollama Cloud API (https://ollama.com/v1/models) returns the same models without those suffixes. In the merge loop:

for m in mdev_models:
    if m and m not in seen:   # "kimi-k2.6:cloud" ≠ "kimi-k2.6" → bypasses dedup
        seen.add(m)
        merged.append(m)       # broken ID added to cache

Users end up with entries like kimi-k2.6:cloud in the model picker. Selecting one triggers a 400/404 from the Ollama API because that ID does not exist.

The fix

Added _strip_ollama_cloud_suffix(model_id) that removes :cloud and -cloud tail suffixes, and applies it to each mdev model ID before the dedup check:

for m in mdev_models:
    normalized = _strip_ollama_cloud_suffix(m)
    if normalized and normalized not in seen:
        seen.add(normalized)
        merged.append(normalized)

Applies to both the case where live API models are available (prevents suffixed duplicates) and the no-API-key fallback (strips suffixes from mdev-only results).

Test plan

• Before: old merge code produces ['kimi-k2.6', 'glm-5.1', 'kimi-k2.6:cloud', 'glm-5.1:cloud', 'qwen3-coder:480b-cloud'] — confirmed locally
• After: test_no_duplicate_when_live_clean_and_mdev_suffixed — exactly one kimi-k2.6, no :cloud entries
• :cloud suffix stripping: test_strips_colon_cloud_suffix
• -cloud suffix stripping: test_strips_dash_cloud_suffix
• Unsuffixed IDs unchanged: test_unsuffixed_model_id_unchanged
• Helper unit test: test_strip_suffix_helper
• All 45 existing + new test_ollama_cloud_provider.py tests pass

Related

• Fixes [Bug]: Ollama Cloud model list shows wrong names with :cloud / -cloud suffix (e.g. kimi-k2.6:cloud), causing 400/404 #16179

🤖 Generated with Claude Code

[Copilot]

Pull request overview

Fixes Ollama Cloud model ID duplication and invalid selections by normalizing models.dev-sourced IDs to match the live Ollama Cloud API, preventing :cloud / -cloud-suffixed entries from leaking into /model and hermes model.

Changes:

• Added _strip_ollama_cloud_suffix() to remove trailing :cloud / -cloud from Ollama Cloud model IDs.
• Applied normalization during the models.dev merge in fetch_ollama_cloud_models() so deduplication works correctly.
• Added targeted unit/integration tests to verify suffix stripping and dedup behavior.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
hermes_cli/models.py	Adds suffix-stripping helper and normalizes models.dev IDs during the merge to prevent duplicates/broken IDs.
tests/hermes_cli/test_ollama_cloud_provider.py	Adds coverage for :cloud / -cloud stripping and confirms no duplicates when live API and models.dev disagree.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[teknium1]

Salvaged via #19881 onto current main - your commit authorship was preserved. Thanks!