vision-agent 0.2.210__py3-none-any.whl → 0.2.211__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- vision_agent/tools/__init__.py +1 -14
- vision_agent/tools/tool_utils.py +2 -2
- vision_agent/tools/tools.py +526 -757
- vision_agent/utils/image_utils.py +16 -0
- {vision_agent-0.2.210.dist-info → vision_agent-0.2.211.dist-info}/METADATA +1 -1
- {vision_agent-0.2.210.dist-info → vision_agent-0.2.211.dist-info}/RECORD +8 -8
- {vision_agent-0.2.210.dist-info → vision_agent-0.2.211.dist-info}/LICENSE +0 -0
- {vision_agent-0.2.210.dist-info → vision_agent-0.2.211.dist-info}/WHEEL +0 -0
vision_agent/tools/__init__.py
CHANGED
@@ -23,40 +23,27 @@ from .tools import (
|
|
23
23
|
TOOLS_INFO,
|
24
24
|
UTIL_TOOLS,
|
25
25
|
UTILITIES_DOCSTRING,
|
26
|
-
blip_image_caption,
|
27
26
|
claude35_text_extraction,
|
28
|
-
clip,
|
29
27
|
closest_box_distance,
|
30
28
|
closest_mask_distance,
|
31
|
-
countgd_example_based_counting,
|
32
29
|
countgd_object_detection,
|
33
30
|
countgd_sam2_object_detection,
|
31
|
+
countgd_example_based_counting,
|
34
32
|
depth_anything_v2,
|
35
33
|
detr_segmentation,
|
36
|
-
dpt_hybrid_midas,
|
37
34
|
extract_frames_and_timestamps,
|
38
|
-
florence2_image_caption,
|
39
35
|
florence2_ocr,
|
40
36
|
florence2_phrase_grounding,
|
41
37
|
florence2_phrase_grounding_video,
|
42
|
-
florence2_roberta_vqa,
|
43
38
|
florence2_sam2_image,
|
44
39
|
florence2_sam2_video_tracking,
|
45
40
|
flux_image_inpainting,
|
46
41
|
generate_pose_image,
|
47
|
-
generate_soft_edge_image,
|
48
42
|
get_tool_documentation,
|
49
43
|
get_tool_recommender,
|
50
|
-
git_vqa_v2,
|
51
44
|
gpt4o_image_vqa,
|
52
45
|
gpt4o_video_vqa,
|
53
|
-
grounding_dino,
|
54
|
-
grounding_sam,
|
55
|
-
ixc25_image_vqa,
|
56
|
-
ixc25_video_vqa,
|
57
46
|
load_image,
|
58
|
-
loca_visual_prompt_counting,
|
59
|
-
loca_zero_shot_counting,
|
60
47
|
minimum_distance,
|
61
48
|
ocr,
|
62
49
|
overlay_bounding_boxes,
|
vision_agent/tools/tool_utils.py
CHANGED
@@ -27,6 +27,7 @@ _LND_API_URL_v2 = f"{_LND_BASE_URL}/v1/tools"
|
|
27
27
|
|
28
28
|
class ToolCallTrace(BaseModel):
|
29
29
|
endpoint_url: str
|
30
|
+
type: str
|
30
31
|
request: MutableMapping[str, Any]
|
31
32
|
response: MutableMapping[str, Any]
|
32
33
|
error: Optional[Error]
|
@@ -221,7 +222,6 @@ def _call_post(
|
|
221
222
|
else:
|
222
223
|
response = session.post(url, json=payload)
|
223
224
|
|
224
|
-
# make sure function_name is in the payload so we can display it
|
225
225
|
tool_call_trace_payload = (
|
226
226
|
payload
|
227
227
|
if "function_name" in payload
|
@@ -229,6 +229,7 @@ def _call_post(
|
|
229
229
|
)
|
230
230
|
tool_call_trace = ToolCallTrace(
|
231
231
|
endpoint_url=url,
|
232
|
+
type="tool_call",
|
232
233
|
request=tool_call_trace_payload,
|
233
234
|
response={},
|
234
235
|
error=None,
|
@@ -252,7 +253,6 @@ def _call_post(
|
|
252
253
|
finally:
|
253
254
|
if tool_call_trace is not None:
|
254
255
|
trace = tool_call_trace.model_dump()
|
255
|
-
trace["type"] = "tool_call"
|
256
256
|
display({MimeType.APPLICATION_JSON: trace}, raw=True)
|
257
257
|
|
258
258
|
|