vision-agent 0.2.145__py3-none-any.whl → 0.2.147__py3-none-any.whl

vision_agent/tools/meta_tools.py

@@ -643,7 +643,23 @@ def use_object_detection_fine_tuning(
         return output_str
 
     code = artifacts[name]
-    patterns = [
+
+    patterns_with_fine_tune_id = [
+        (
+            r'florence2_phrase_grounding\(\s*"([^"]+)"\s*,\s*([^,]+)(?:,\s*"[^"]+")?\s*\)',
+            lambda match: f'florence2_phrase_grounding("{match.group(1)}", {match.group(2)}, "{fine_tune_id}")',
+        ),
+        (
+            r'owl_v2_image\(\s*"([^"]+)"\s*,\s*([^,]+)(?:,\s*"[^"]+")?\s*\)',
+            lambda match: f'owl_v2_image("{match.group(1)}", {match.group(2)}, "{fine_tune_id}")',
+        ),
+        (
+            r'florence2_sam2_image\(\s*"([^"]+)"\s*,\s*([^,]+)(?:,\s*"[^"]+")?\s*\)',
+            lambda match: f'florence2_sam2_image("{match.group(1)}", {match.group(2)}, "{fine_tune_id}")',
+        ),
+    ]
+
+    patterns_without_fine_tune_id = [
         (
             r"florence2_phrase_grounding\(\s*([^\)]+)\s*\)",
             lambda match: f'florence2_phrase_grounding({match.group(1)}, "{fine_tune_id}")',
@@ -659,8 +675,21 @@ def use_object_detection_fine_tuning(
     ]
 
     new_code = code
-    for pattern, replacer in patterns:
-        new_code = re.sub(pattern, replacer, new_code)
+
+    for index, (pattern_with_fine_tune_id, replacer_with_fine_tune_id) in enumerate(
+        patterns_with_fine_tune_id
+    ):
+        if re.search(pattern_with_fine_tune_id, new_code):
+            new_code = re.sub(
+                pattern_with_fine_tune_id, replacer_with_fine_tune_id, new_code
+            )
+        else:
+            (pattern_without_fine_tune_id, replacer_without_fine_tune_id) = (
+                patterns_without_fine_tune_id[index]
+            )
+            new_code = re.sub(
+                pattern_without_fine_tune_id, replacer_without_fine_tune_id, new_code
+            )
 
     if new_code == code:
         output_str = (

vision_agent/tools/tools.py

@@ -1759,14 +1759,17 @@ def _save_video_to_result(video_uri: str) -> None:
 
 
 def overlay_bounding_boxes(
-    image: np.ndarray, bboxes: List[Dict[str, Any]]
-) -> np.ndarray:
+    medias: Union[np.ndarray, List[np.ndarray]],
+    bboxes: Union[List[Dict[str, Any]], List[List[Dict[str, Any]]]],
+) -> Union[np.ndarray, List[np.ndarray]]:
     """'overlay_bounding_boxes' is a utility function that displays bounding boxes on
     an image.
 
     Parameters:
-        image (np.ndarray): The image to display the bounding boxes on.
-        bboxes (List[Dict[str, Any]]): A list of dictionaries containing the bounding
+        medias (Union[np.ndarray, List[np.ndarra]]): The image or frames to display the
+            bounding boxes on.
+        bboxes (Union[List[Dict[str, Any]], List[List[Dict[str, Any]]]]): A list of
+            dictionaries or a list of list of dictionaries containing the bounding
             boxes.
 
     Returns:
@@ -1778,41 +1781,54 @@ def overlay_bounding_boxes(
             image, [{'score': 0.99, 'label': 'dinosaur', 'bbox': [0.1, 0.11, 0.35, 0.4]}],
         )
     """
-    pil_image = Image.fromarray(image.astype(np.uint8)).convert("RGB")
 
-    if len(set([box["label"] for box in bboxes])) > len(COLORS):
+    medias_int: List[np.ndarray] = (
+        [medias] if isinstance(medias, np.ndarray) else medias
+    )
+    bbox_int = [bboxes] if isinstance(bboxes[0], dict) else bboxes
+    bbox_int = cast(List[List[Dict[str, Any]]], bbox_int)
+    labels = set([bb["label"] for b in bbox_int for bb in b])
+
+    if len(labels) > len(COLORS):
         _LOGGER.warning(
             "Number of unique labels exceeds the number of available colors. Some labels may have the same color."
         )
 
-    color = {
-        label: COLORS[i % len(COLORS)]
-        for i, label in enumerate(set([box["label"] for box in bboxes]))
-    }
-    bboxes = sorted(bboxes, key=lambda x: x["label"], reverse=True)
+    color = {label: COLORS[i % len(COLORS)] for i, label in enumerate(labels)}
 
-    width, height = pil_image.size
-    fontsize = max(12, int(min(width, height) / 40))
-    draw = ImageDraw.Draw(pil_image)
-    font = ImageFont.truetype(
-        str(resources.files("vision_agent.fonts").joinpath("default_font_ch_en.ttf")),
-        fontsize,
-    )
+    frame_out = []
+    for i, frame in enumerate(medias_int):
+        pil_image = Image.fromarray(frame.astype(np.uint8)).convert("RGB")
 
-    for elt in bboxes:
-        label = elt["label"]
-        box = elt["bbox"]
-        scores = elt["score"]
+        bboxes = bbox_int[i]
+        bboxes = sorted(bboxes, key=lambda x: x["label"], reverse=True)
 
-        # denormalize the box if it is normalized
-        box = denormalize_bbox(box, (height, width))
+        width, height = pil_image.size
+        fontsize = max(12, int(min(width, height) / 40))
+        draw = ImageDraw.Draw(pil_image)
+        font = ImageFont.truetype(
+            str(
+                resources.files("vision_agent.fonts").joinpath("default_font_ch_en.ttf")
+            ),
+            fontsize,
+        )
 
-        draw.rectangle(box, outline=color[label], width=4)
-        text = f"{label}: {scores:.2f}"
-        text_box = draw.textbbox((box[0], box[1]), text=text, font=font)
-        draw.rectangle((box[0], box[1], text_box[2], text_box[3]), fill=color[label])
-        draw.text((box[0], box[1]), text, fill="black", font=font)
-    return np.array(pil_image)
+        for elt in bboxes:
+            label = elt["label"]
+            box = elt["bbox"]
+            scores = elt["score"]
+
+            # denormalize the box if it is normalized
+            box = denormalize_bbox(box, (height, width))
+            draw.rectangle(box, outline=color[label], width=4)
+            text = f"{label}: {scores:.2f}"
+            text_box = draw.textbbox((box[0], box[1]), text=text, font=font)
+            draw.rectangle(
+                (box[0], box[1], text_box[2], text_box[3]), fill=color[label]
+            )
+            draw.text((box[0], box[1]), text, fill="black", font=font)
+        frame_out.append(np.array(pil_image))
+    return frame_out[0] if len(frame_out) == 1 else frame_out
 
 
 def _get_text_coords_from_mask(
@@ -1852,7 +1868,8 @@ def overlay_segmentation_masks(
         medias (Union[np.ndarray, List[np.ndarray]]): The image or frames to display
             the masks on.
         masks (Union[List[Dict[str, Any]], List[List[Dict[str, Any]]]]): A list of
-            dictionaries containing the masks, labels and scores.
+            dictionaries or a list of list of dictionaries containing the masks, labels
+            and scores.
         draw_label (bool, optional): If True, the labels will be displayed on the image.
         secondary_label_key (str, optional): The key to use for the secondary
             tracking label which is needed in videos to display tracking information.

{vision_agent-0.2.145.dist-info → vision_agent-0.2.147.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vision-agent
-Version: 0.2.145
+Version: 0.2.147
 Summary: Toolset for Vision Agent
 Author: Landing AI
 Author-email: dev@landing.ai

{vision_agent-0.2.145.dist-info → vision_agent-0.2.147.dist-info}/RECORD

@@ -15,10 +15,10 @@ vision_agent/lmm/__init__.py,sha256=jyY1sJb_tYKg5-Wzs3p1lvwFkc-aUNZfMcLy3TOC4Zg,
 vision_agent/lmm/lmm.py,sha256=B5ClgwvbybVCWkf9opDMLjTtJZemUU4KUkQoRxGh43I,16787
 vision_agent/lmm/types.py,sha256=ZEXR_ptBL0ZwDMTDYkgxUCmSZFmBYPQd2jreNzr_8UY,221
 vision_agent/tools/__init__.py,sha256=zUv3aVPN1MXfyQiQi5To4rkQGtG7mxLQ1NjLI3pxM80,2412
-vision_agent/tools/meta_tools.py,sha256=iHvMeBktWcVi-0DOrSMak1gsZrM_VKJlAq1mAFbBemE,23477
+vision_agent/tools/meta_tools.py,sha256=skY4nHSH7PFXNBAEKGYI6XZskv8zJ2UX1KbMY2gd_bo,24687
 vision_agent/tools/prompts.py,sha256=V1z4YJLXZuUl_iZ5rY0M5hHc_2tmMEUKr0WocXKGt4E,1430
 vision_agent/tools/tool_utils.py,sha256=5ukuDMxbEH4iKetYR9I7twzsA8ECyP4tVwYXQq54mxI,8020
-vision_agent/tools/tools.py,sha256=lLGZ-GgORSbFCTaBWI07Om1L8gnO00d8MN600YC7s_o,73773
+vision_agent/tools/tools.py,sha256=c7SjtZD7YfxhEAGYYe-ExVCBA4NDXmRwerBIbd-XEH8,74557
 vision_agent/tools/tools_types.py,sha256=JUOZWGW2q-dlJ85CHr9gvo9KQk_rXyjJhi-iwPNn4eM,2397
 vision_agent/utils/__init__.py,sha256=7fMgbZiEwbNS0fBOS_hJI5PuEYBblw36zLi_UjUzvj4,244
 vision_agent/utils/exceptions.py,sha256=booSPSuoULF7OXRr_YbC4dtKt6gM_HyiFQHBuaW86C4,2052
@@ -27,7 +27,7 @@ vision_agent/utils/image_utils.py,sha256=rm9GfXvD4JrjnqKrP_f2gfq4SzmqYC0IdC1kKwd
 vision_agent/utils/sim.py,sha256=ebE9Cs00pVEDI1HMjAzUBk88tQQmc2U-yAzIDinnekU,5572
 vision_agent/utils/type_defs.py,sha256=BE12s3JNQy36QvauXHjwyeffVh5enfcvd4vTzSwvEZI,1384
 vision_agent/utils/video.py,sha256=xbMEoRk13l4fHeQlbvMQhLCn8RNndYmsDhUf01TUeR8,4781
-vision_agent-0.2.145.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-vision_agent-0.2.145.dist-info/METADATA,sha256=Z7gVyhZAtrpb3zTrFl2LZli9hw6f2VW_0p5_MQtpURM,13758
-vision_agent-0.2.145.dist-info/WHEEL,sha256=7Z8_27uaHI_UZAc4Uox4PpBhQ9Y5_modZXWMxtUi4NU,88
-vision_agent-0.2.145.dist-info/RECORD,,
+vision_agent-0.2.147.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+vision_agent-0.2.147.dist-info/METADATA,sha256=juNWGPjUPW3pxyGKPEBV_CwDllMnTj8B6c9Z7hevCyM,13758
+vision_agent-0.2.147.dist-info/WHEEL,sha256=7Z8_27uaHI_UZAc4Uox4PpBhQ9Y5_modZXWMxtUi4NU,88
+vision_agent-0.2.147.dist-info/RECORD,,