ling/hermes-agent
1
0
Fork 0
mirror of https://github.com/NousResearch/hermes-agent.git synced 2026-04-28 06:51:16 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix(anthropic): revert inline vision, add hermes model flow, wire vision aux

Browse Source 
Feedback fixes:

1. Revert _convert_vision_content — vision is handled by the vision_analyze
   tool, not by converting image blocks inline in conversation messages.
   Removed the function and its tests.

2. Add Anthropic to 'hermes model' (cmd_model in main.py):
   - Added to provider_labels dict
   - Added to providers selection list
   - Added _model_flow_anthropic() with Claude Code credential auto-detection,
     API key prompting, and model selection from catalog.

3. Wire up Anthropic as a vision-capable auxiliary provider:
   - Added _try_anthropic() to auxiliary_client.py using claude-sonnet-4
     as the vision model (Claude natively supports multimodal)
   - Added to the get_vision_auxiliary_client() auto-detection chain
     (after OpenRouter/Nous, before Codex/custom)

Cache tracking note: the Anthropic cache metrics branch in run_agent.py
(cache_read_input_tokens / cache_creation_input_tokens) is in the correct
place — it's response-level parsing, same location as the existing
OpenRouter cache tracking. auxiliary_client.py has no cache tracking.
This commit is contained in:
teknium1
2026-03-12 16:09:04 -07:00
parent d7adfe8f61
commit 7086fde37e
4 changed files with 105 additions and 94 deletions
Download Patch File Download Diff File Expand all files Collapse all files

37
tests/test_anthropic_adapter.py
View File

@@ -413,43 +413,6 @@ class TestNormalizeResponse:
assert len(msg.tool_calls) == 1
# ---------------------------------------------------------------------------
# Vision content conversion
# ---------------------------------------------------------------------------
class TestVisionContentConversion:
def test_base64_image(self):
from agent.anthropic_adapter import _convert_vision_content
content = [
{"type": "text", "text": "What's in this image?"},
{"type": "image_url", "image_url": {"url": "data:image/png;base64,iVBOR"}},
]
result = _convert_vision_content(content)
assert result[0] == {"type": "text", "text": "What's in this image?"}
assert result[1]["type"] == "image"
assert result[1]["source"]["type"] == "base64"
assert result[1]["source"]["media_type"] == "image/png"
assert result[1]["source"]["data"] == "iVBOR"
def test_url_image(self):
from agent.anthropic_adapter import _convert_vision_content
content = [
{"type": "image_url", "image_url": {"url": "https://example.com/img.png"}},
]
result = _convert_vision_content(content)
assert result[0]["type"] == "image"
assert result[0]["source"]["type"] == "url"
assert result[0]["source"]["url"] == "https://example.com/img.png"
def test_passthrough_non_list(self):
from agent.anthropic_adapter import _convert_vision_content
assert _convert_vision_content("plain text") == "plain text"
# ---------------------------------------------------------------------------
# Role alternation
# ---------------------------------------------------------------------------