descriptron-sam3 2.5.1__py3-none-any.whl

descriptron_sam3/__init__.py

@@ -0,0 +1,2 @@
+"""descriptron-sam3: SAM 3 multi-instance proposals for Descriptron (see descriptron-sam3 --help)."""
+__version__ = "2.5.1"

descriptron_sam3/cli.py

@@ -0,0 +1,15 @@
+"""descriptron_sam3.cli - run descriptron_sam3_instances.py (shipped in tools/, with SAM 3's text vocabulary)."""
+from __future__ import annotations
+
+import runpy
+import sys
+from importlib.resources import files
+from pathlib import Path
+
+
+def main() -> None:
+    path = Path(str(files("descriptron_sam3") / "tools" / "descriptron_sam3_instances.py"))
+    if not path.is_file():
+        raise SystemExit(f"{path} is missing - was sync_sources.py run before building?")
+    sys.argv = [str(path), *sys.argv[1:]]
+    runpy.run_path(str(path), run_name="__main__")

descriptron_sam3/tools/descriptron_sam3_instances.py

@@ -0,0 +1,271 @@
+#!/usr/bin/env python3
+"""descriptron_sam3_instances.py - find every instance of a structure with SAM 3 (multi-instance prediction).
+
+Give SAM 3 a few outlined examples (for example five setae on a wing) and/or a text prompt ("seta"); it proposes
+every matching instance in each image. The result is a COCO file Descriptron opens like any other prediction, so
+the taxonomist deletes false ones and adds misses; corrected files can then train the Mask R-CNN detector.
+
+Small, thin structures disappear when a whole photograph is shrunk to SAM 3's working size (1008 px), so every
+image is cut into overlapping tiles at its own resolution, each tile is prompted, masks are mapped back to the
+full image, and duplicates from overlapping tiles are merged (box IoU).
+
+Prompts per tile:
+  --text "seta"               text prompt (all tiles)
+  --exemplars examples.json   COCO file with outlined examples of --category; each example box is a positive
+                              visual prompt on the tiles that contain it (SAM 3 box prompts work within an image)
+  --boxes / --points          prompts drawn in the GUI on ONE image (x0,y0,x1,y1;... and x,y,label;...): each
+                              box is an example; clicked points first outline one example (SAM 3's SAM-1-style
+                              interactive mode), whose box then becomes an example
+  --tile 0                    no tiling: the whole image at SAM 3's working size
+Runs in its own environment (Python 3.12, PyTorch >= 2.7): Descriptron's `sam3` conda env. The checkpoints are
+gated: request access at https://huggingface.co/facebook/sam3 and log in (`hf auth login`) once.
+
+  python descriptron_sam3_instances.py --images wings/ --text seta --category seta --out setae_sam3.json
+  python descriptron_sam3_instances.py --images wings/ --exemplars five_setae.json --category seta \\
+      --text seta --tile 1008 --overlap 0.25 --confidence 0.5 --out setae_sam3.json
+"""
+import argparse, json, os, sys, time
+from pathlib import Path
+import numpy as np
+
+IMG_EXT = {".png", ".jpg", ".jpeg", ".tif", ".tiff", ".bmp"}
+
+
+def parse_boxes(text):
+    """'x0,y0,x1,y1;x0,y0,x1,y1' -> [[x0, y0, x1, y1], ...] (corners in any order)."""
+    out = []
+    for part in (text or "").split(";"):
+        if part.strip():
+            x0, y0, x1, y1 = (float(v) for v in part.split(","))
+            out.append([min(x0, x1), min(y0, y1), max(x0, x1), max(y0, y1)])
+    return out
+
+
+def parse_points(text):
+    """'x,y,label;...' -> ([[x, y], ...], [label, ...]); label 1 = on the structure, 0 = background."""
+    pts, labs = [], []
+    for part in (text or "").split(";"):
+        if part.strip():
+            x, y, lab = part.split(","); pts.append([float(x), float(y)]); labs.append(int(float(lab)))
+    return pts, labs
+
+
+def point_window(pts, w, h, size):
+    """a window at native resolution around clicked points: at least size x size (smaller if the image is), centred
+    on the points, inside the image. Thin structures vanish when a big photo is shrunk to SAM 3's 1008 px."""
+    xs = [p[0] for p in pts]; ys = [p[1] for p in pts]
+    x0, x1, y0, y1 = min(xs), max(xs), min(ys), max(ys)
+    ww = min(w, max(size, x1 - x0 + size // 4)); hh = min(h, max(size, y1 - y0 + size // 4))
+    cx, cy = (x0 + x1) / 2, (y0 + y1) / 2
+    left = int(min(max(0, cx - ww / 2), w - ww)); top = int(min(max(0, cy - hh / 2), h - hh))
+    return left, top, int(ww), int(hh)
+
+
+def outline_points(proc, model, im, pts, labs, size):
+    """SAM 3 point outline (SAM-1 style) on a native-resolution window; returns (full-image mask, score)."""
+    W, H = im.size
+    pos = [p for p, l in zip(pts, labs) if l == 1] or pts
+    x0, y0, ww, hh = point_window(pos, W, H, size)
+    keep = [(p, l) for p, l in zip(pts, labs) if x0 <= p[0] < x0 + ww and y0 <= p[1] < y0 + hh]
+    st = proc.set_image(im.crop((x0, y0, x0 + ww, y0 + hh)))
+    pm, ps, _ = model.predict_inst(st, point_coords=np.array([[p[0] - x0, p[1] - y0] for p, _ in keep], float),
+                                   point_labels=np.array([l for _, l in keep]), multimask_output=False)
+    full = np.zeros((H, W), bool)
+    m = np.squeeze(np.asarray(pm)[0])[:hh, :ww] > 0
+    if m.sum() > 0.5 * ww * hh:              # the click missed the structure: SAM returns the background
+        print(f"point {pos[0]}: outline covers most of its window (click on background?) - skipped", flush=True)
+        return full, 0.0
+    full[y0:y0 + hh, x0:x0 + ww] = m
+    return full, float(np.asarray(ps).ravel()[0])
+
+
+def find_bpe_vocab():
+    """SAM 3's text-encoder vocabulary: sam3's own copy (git install), else the copy shipped with Descriptron
+    (the sam3 0.1.4 wheel on PyPI leaves it out)."""
+    name = "bpe_simple_vocab_16e6.txt.gz"
+    cands = []
+    try:
+        import sam3
+        d = os.path.dirname(sam3.__file__)
+        cands += [os.path.join(d, "assets", name), os.path.join(d, "..", "assets", name)]
+    except Exception:
+        pass
+    here = os.path.dirname(os.path.abspath(__file__))
+    cands += [os.path.join(here, "sam3_assets", name), os.path.join(here, "..", "sam3_assets", name)]
+    for c in cands:
+        if os.path.isfile(c):
+            return os.path.abspath(c)
+    return None
+
+
+def tiles(w, h, size, overlap):
+    """top-left corners of overlapping tiles covering a w x h image (one tile if the image is small; size 0 = one)."""
+    if size <= 0 or (w <= size and h <= size):
+        return [(0, 0, w, h)]
+    step = max(1, int(size * (1 - overlap)))
+    xs = list(range(0, max(w - size, 0) + 1, step)); ys = list(range(0, max(h - size, 0) + 1, step))
+    if xs[-1] + size < w: xs.append(w - size)
+    if ys[-1] + size < h: ys.append(h - size)
+    return [(x, y, min(size, w), min(size, h)) for y in ys for x in xs]
+
+
+def box_iou(a, b):
+    ix = max(0, min(a[2], b[2]) - max(a[0], b[0])); iy = max(0, min(a[3], b[3]) - max(a[1], b[1]))
+    inter = ix * iy; ua = (a[2] - a[0]) * (a[3] - a[1]) + (b[2] - b[0]) * (b[3] - b[1]) - inter
+    return inter / ua if ua > 0 else 0.0
+
+
+def merge(dets, iou):
+    """greedy non-maximum suppression across tiles (keeps the higher-scoring of two overlapping detections)."""
+    dets = sorted(dets, key=lambda d: -d["score"]); kept = []
+    for d in dets:
+        if all(box_iou(d["box"], k["box"]) < iou for k in kept):
+            kept.append(d)
+    return kept
+
+
+def mask_to_polygons(mask):
+    import cv2
+    cs, _ = cv2.findContours(mask.astype(np.uint8), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
+    return [c.flatten().astype(float).tolist() for c in cs if len(c) >= 3]
+
+
+def load_exemplars(path, category):
+    """example boxes [x0, y0, x1, y1] per image file name, from a COCO file (annotations of `category`)."""
+    d = json.load(open(path)); cats = {c["id"]: c["name"] for c in d["categories"]}
+    names = {im["id"]: os.path.basename(im["file_name"]) for im in d["images"]}
+    out = {}
+    for a in d["annotations"]:
+        if category and cats.get(a["category_id"]) != category:
+            continue
+        x, y, w, h = a["bbox"]; out.setdefault(names[a["image_id"]], []).append([x, y, x + w, y + h])
+    return out
+
+
+def main(argv=None):
+    ap = argparse.ArgumentParser(description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter)
+    ap.add_argument("--images", required=True, help="folder of images (or one image file)")
+    ap.add_argument("--text", default=None, help="text prompt, e.g. seta")
+    ap.add_argument("--exemplars", default=None, help="COCO file with a few outlined examples (positive box prompts)")
+    ap.add_argument("--boxes", default=None, help="example boxes on a single image: x0,y0,x1,y1;...")
+    ap.add_argument("--points", default=None, help="clicked points on a single image: x,y,label;... (outline one example)")
+    ap.add_argument("--points_mode", default="example", choices=["example", "each", "examples"],
+                    help="example: all points outline ONE example and SAM 3 finds all like it; each: one mask per "
+                         "positive point (negative points shared), no search; examples: one outline per positive "
+                         "point, then all outlines together are the examples for the search")
+    ap.add_argument("--point_window", type=int, default=1008,
+                    help="clicked points are outlined on a window of this many native pixels around them (default 1008)")
+    ap.add_argument("--category", default=None, help="category name to write (and to read from --exemplars)")
+    ap.add_argument("--tile", type=int, default=1008, help="tile size in pixels (SAM 3 works at 1008; 0 = whole image, no tiling)")
+    ap.add_argument("--overlap", type=float, default=0.25, help="tile overlap fraction (default 0.25)")
+    ap.add_argument("--confidence", type=float, default=0.5, help="keep detections at or above this score")
+    ap.add_argument("--merge_iou", type=float, default=0.5, help="box IoU above which tile duplicates merge")
+    ap.add_argument("--min_area", type=int, default=4, help="drop masks smaller than this many pixels")
+    ap.add_argument("--device", default="cuda")
+    ap.add_argument("--out", required=True, help="output COCO json")
+    a = ap.parse_args(argv)
+    if not (a.text or a.exemplars or a.boxes or a.points):
+        ap.error("give --text, --exemplars, --boxes and/or --points")
+    if a.points_mode in ("each", "examples") and not a.points:
+        ap.error("--points_mode each needs --points")
+    if (a.boxes or a.points) and not Path(a.images).is_file():
+        ap.error("--boxes / --points are drawn on one image: give that image file as --images")
+    cat_name = a.category or a.text or "instance"
+    from PIL import Image
+    import torch
+    from sam3.model_builder import build_sam3_image_model
+    from sam3.model.sam3_image_processor import Sam3Processor
+    try:
+        model = build_sam3_image_model(bpe_path=find_bpe_vocab(), enable_inst_interactivity=bool(a.points))
+    except Exception as e:
+        sys.exit(f"Could not load SAM 3 ({e}).\nThe checkpoints are gated: request access at "
+                 "https://huggingface.co/facebook/sam3, then run `hf auth login` once in the sam3 environment.")
+    proc = Sam3Processor(model, resolution=a.tile if a.tile > 0 else 1008, device=a.device,
+                         confidence_threshold=a.confidence)
+    ex = load_exemplars(a.exemplars, a.category) if a.exemplars else {}
+    p = Path(a.images); files = [p] if p.is_file() else sorted(f for f in p.iterdir() if f.suffix.lower() in IMG_EXT)
+    coco = {"images": [], "annotations": [], "categories": [{"id": 1, "name": cat_name}],
+            "info": {"description": "SAM 3 multi-instance proposals (descriptron_sam3_instances.py); review before use",
+                     "text_prompt": a.text, "exemplars": a.exemplars, "tile": a.tile, "overlap": a.overlap,
+                     "confidence": a.confidence}}
+    t0 = time.time(); aid = 0
+    for k, f in enumerate(files, 1):
+        im = Image.open(f).convert("RGB"); W, H = im.size
+        coco["images"].append({"id": k, "file_name": f.name, "width": W, "height": H})
+        boxes_here = ex.get(f.name, []) + parse_boxes(a.boxes)
+        dets = []
+        if a.points and a.points_mode in ("each", "examples"):   # one structure per positive click
+            pts, labs = parse_points(a.points)
+            neg = [q for q, l in zip(pts, labs) if l == 0]
+            with torch.inference_mode(), torch.autocast("cuda", dtype=torch.bfloat16, enabled=a.device == "cuda"):
+                for q, l in zip(pts, labs):
+                    if l != 1:
+                        continue
+                    m, sc = outline_points(proc, model, im, [q] + neg, [1] + [0] * len(neg), a.point_window)
+                    if m.sum() > 0.5 * W * H:              # a click on the background returns the background
+                        print(f"point {q}: mask covers most of the image (background?) - skipped", flush=True)
+                    elif m.sum() >= a.min_area:
+                        ys_, xs_ = np.nonzero(m)
+                        dets.append({"mask": m, "score": sc,
+                                     "box": [float(xs_.min()), float(ys_.min()), float(xs_.max() + 1), float(ys_.max() + 1)]})
+            if a.points_mode == "each":                # nothing else
+                boxes_here = []
+                tile_list = []
+            else:                                      # "examples": every outline's box is an example, together
+                boxes_here += [d["box"] for d in dets]
+                print(f"{len(dets)} clicked outline(s) -> examples", flush=True)
+                tile_list = tiles(W, H, a.tile, a.overlap)
+        else:
+            tile_list = tiles(W, H, a.tile, a.overlap)
+        if a.points and a.points_mode == "example":   # clicked points -> one outlined example (whole image) -> its box
+            pts, labs = parse_points(a.points)
+            with torch.inference_mode(), torch.autocast("cuda", dtype=torch.bfloat16, enabled=a.device == "cuda"):
+                m, _sc = outline_points(proc, model, im, pts, labs, a.point_window)
+            if m.any():
+                ys_, xs_ = np.nonzero(m)
+                boxes_here.append([float(xs_.min()), float(ys_.min()), float(xs_.max() + 1), float(ys_.max() + 1)])
+                print(f"points -> example box {boxes_here[-1]}", flush=True)
+            else:
+                print("points gave no example mask", flush=True)
+        for (x0, y0, tw, th) in tile_list:
+            tile = im.crop((x0, y0, x0 + tw, y0 + th))
+            with torch.inference_mode(), torch.autocast("cuda", dtype=torch.bfloat16, enabled=a.device == "cuda"):
+                st = proc.set_image(tile)
+                out = None
+                if a.text:
+                    out = proc.set_text_prompt(prompt=a.text, state=st)
+                for (bx0, by0, bx1, by1) in boxes_here:             # examples inside this tile only
+                    cx0, cy0, cx1, cy1 = max(bx0, x0), max(by0, y0), min(bx1, x0 + tw), min(by1, y0 + th)
+                    if cx1 - cx0 < 2 or cy1 - cy0 < 2:
+                        continue
+                    nb = [((cx0 + cx1) / 2 - x0) / tw, ((cy0 + cy1) / 2 - y0) / th, (cx1 - cx0) / tw, (cy1 - cy0) / th]
+                    out = proc.add_geometric_prompt(box=nb, label=True, state=st)
+                if out is None:
+                    continue
+            masks = out["masks"].float().cpu().numpy(); scores = out["scores"].float().cpu().numpy()
+            boxes = out["boxes"].float().cpu().numpy()
+            for m, s, b in zip(masks, scores, boxes):
+                m = np.squeeze(m) > 0.5
+                if m.sum() < a.min_area:
+                    continue
+                full = np.zeros((H, W), bool); full[y0:y0 + th, x0:x0 + tw] = m[:th, :tw]
+                dets.append({"mask": full, "score": float(s),
+                             "box": [float(b[0] + x0), float(b[1] + y0), float(b[2] + x0), float(b[3] + y0)]})
+        kept = dets if a.points_mode == "each" and a.points else merge(dets, a.merge_iou)
+        for dd in kept:
+            polys = mask_to_polygons(dd["mask"])
+            if not polys:
+                continue
+            ys_, xs_ = np.nonzero(dd["mask"]); aid += 1
+            coco["annotations"].append({"id": aid, "image_id": k, "category_id": 1, "segmentation": polys,
+                                        "area": float(dd["mask"].sum()), "iscrowd": 0, "score": dd["score"],
+                                        "bbox": [float(xs_.min()), float(ys_.min()), float(xs_.max() - xs_.min() + 1),
+                                                 float(ys_.max() - ys_.min() + 1)]})
+        print(f"[{k}/{len(files)}] {f.name}: {len(kept)} instances", flush=True)
+    Path(a.out).parent.mkdir(parents=True, exist_ok=True)
+    json.dump(coco, open(a.out, "w"))
+    print(f"wrote {a.out}: {aid} instances in {len(files)} images ({time.time() - t0:.0f} s)")
+
+
+if __name__ == "__main__":
+    main()

descriptron_sam3/tools/sam3_assets/README.txt

@@ -0,0 +1,4 @@
+bpe_simple_vocab_16e6.txt.gz - the byte-pair-encoding vocabulary of OpenAI CLIP (MIT licence,
+https://github.com/openai/CLIP), used by SAM 3's text encoder. The sam3 wheel on PyPI (0.1.4) does not include it,
+so Descriptron ships this copy; descriptron_sam3_instances.py uses sam3's own copy when there is one.
+md5 933b7abbbbde62c36f02f0e6ccde464f

descriptron_sam3-2.5.1.dist-info/METADATA

@@ -0,0 +1,53 @@
+Metadata-Version: 2.5
+Name: descriptron-sam3
+Version: 2.5.1
+Summary: SAM 3 for Descriptron: find every instance of a small structure (e.g. setae) from examples, clicks or a word
+Project-URL: Homepage, https://github.com/alexrvandam/Descriptron
+Author: Alex R. Van Dam
+License: Apache-2.0
+License-File: LICENSE
+License-File: NOTICE
+Classifier: Intended Audience :: Science/Research
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: Programming Language :: Python :: 3
+Classifier: Topic :: Scientific/Engineering :: Image Recognition
+Requires-Python: >=3.12
+Requires-Dist: einops
+Requires-Dist: huggingface-hub
+Requires-Dist: numpy<2,>=1.26
+Requires-Dist: opencv-python-headless>=4.8
+Requires-Dist: pillow>=9
+Requires-Dist: psutil
+Requires-Dist: pycocotools
+Requires-Dist: sam3==0.1.4
+Requires-Dist: setuptools<81
+Requires-Dist: torch>=2.7
+Requires-Dist: torchvision
+Description-Content-Type: text/markdown
+
+# descriptron-sam3
+
+SAM 3 for [Descriptron](https://github.com/alexrvandam/Descriptron): outline one or a few examples of a small
+structure (for example setae on a wing or genitalia), click on them, or type a word, and SAM 3 proposes every
+matching instance as a COCO file. The Descriptron GUI (descriptron-gui 2.5.1 and later) runs it from the dark
+green **Find all instances (SAM 3)** button and shows the instances together, one colour each, to be labelled in
+one step.
+
+SAM 3 needs Python 3.12 and PyTorch >= 2.7, so install it in its own environment. pipx does that and puts the
+`descriptron-sam3` command on your PATH, where the GUI finds it:
+
+    pipx install --python python3.12 descriptron-sam3
+
+The SAM 3 checkpoints are gated: request access at https://huggingface.co/facebook/sam3 (SAM License), then log in
+once (`pipx run --spec huggingface_hub hf auth login`, or `hf auth login` in the environment). The first run
+downloads the checkpoint (3.3 GB). A CUDA GPU is strongly recommended.
+
+    descriptron-sam3 --images wings/ --text bristle --category seta --out setae.json
+    descriptron-sam3 --images wing1.tif --boxes "1200,800,1260,1100" --out setae.json
+    descriptron-sam3 --images wing1.tif --points "1520,4040,1" --points_mode each --tile 0 --out one_seta.json
+
+Results are proposals to review in the GUI, not finished annotations. Examples work within the image they are
+drawn on; a word searches every image tile by tile. If the GUI cannot find the command, point it at the Python of
+the environment that has descriptron-sam3: `DESCRIPTRON_SAM3_PYTHON=/path/to/python`.
+
+Licence: Apache-2.0 (see NOTICE for the CLIP vocabulary file and for SAM 3's own licence).

descriptron_sam3-2.5.1.dist-info/RECORD

@@ -0,0 +1,11 @@
+descriptron_sam3/__init__.py,sha256=S88-3kIeL_TfMeFI7wAfhNgGuIxrCthFdt4uzgS5cH8,124
+descriptron_sam3/cli.py,sha256=CNGqhoH-8QPqGBzc3Kv6gnlIoQ-odYzc6yqeDoHcsZs,561
+descriptron_sam3/tools/descriptron_sam3_instances.py,sha256=k1nmf4nvzypgqHPGjRueYo8OjtAbCkFqpNCL1Kh5jKg,15923
+descriptron_sam3/tools/sam3_assets/README.txt,sha256=NMGYRP3H_cqjE6GJqDlYKBSvRx2ZitufZre_rOWV9fo,348
+descriptron_sam3/tools/sam3_assets/bpe_simple_vocab_16e6.txt.gz,sha256=kkaRrCiOVECSNhFWUq1KolD0ggPeUKnkcipuzUjWgEo,1356917
+descriptron_sam3-2.5.1.dist-info/METADATA,sha256=f1UkuzxFtjWbguXvZXquu9BYQqhfcbrX7qtcfUcPEdM,2617
+descriptron_sam3-2.5.1.dist-info/WHEEL,sha256=W3fkpkm7-wf9vBI5Z-7s0eWkeM-spu78I8Neb98DeEg,87
+descriptron_sam3-2.5.1.dist-info/entry_points.txt,sha256=fCu14JptpZv15TU7m065PHjRSv5dh5oZ4YvCdSZ53jc,63
+descriptron_sam3-2.5.1.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+descriptron_sam3-2.5.1.dist-info/licenses/NOTICE,sha256=RfTMWUQnFbPv_aHlOU_HU3SlTiBwp3WsLw3yfqZG8JQ,2935
+descriptron_sam3-2.5.1.dist-info/RECORD,,

descriptron_sam3-2.5.1.dist-info/WHEEL

@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: hatchling 1.32.4
+Root-Is-Purelib: true
+Tag: py3-none-any

descriptron_sam3-2.5.1.dist-info/entry_points.txt

@@ -0,0 +1,2 @@
+[console_scripts]
+descriptron-sam3 = descriptron_sam3.cli:main

descriptron_sam3-2.5.1.dist-info/licenses/LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

descriptron_sam3-2.5.1.dist-info/licenses/NOTICE

@@ -0,0 +1,69 @@
+Descriptron
+Copyright 2024-2026 Alex R. Van Dam (ORCID 0000-0002-1966-0338)
+
+This product includes software developed by Alex R. Van Dam
+https://github.com/alexrvandam/Descriptron
+
+Licensed under the Apache License, Version 2.0 (see LICENSE). Under Section
+4(d) of that licence, this NOTICE file must be included with any
+redistribution of this software or of any work derived from it.
+
+REQUIRED CITATION
+-----------------
+Anyone who uses Descriptron, or any software derived from it (including
+descriptron-core, descriptron-vision, descriptron-gui and descriptron-mcp), in
+work that is published, presented or distributed must cite:
+
+  1. The software (Descriptron v2), by the DOI of the version used:
+     Van Dam, A. R. (2026). Descriptron: morphology-driven species descriptions,
+     keys and delimitation for dark taxa (Version 2.1.1). Zenodo.
+     https://doi.org/10.5281/zenodo.22958109
+     (all versions: https://doi.org/10.5281/zenodo.17077224;
+      repository: https://github.com/alexrvandam/Descriptron)
+
+  2. The first Descriptron paper:
+     Van Dam, A. R. & Štarhová Serbina, L. (2026). Descriptron: Artificial
+     intelligence for automating taxonomic species descriptions with a
+     user-friendly software package. Systematic Entomology, 51(1), e70005.
+     https://doi.org/10.1111/syen.70005
+
+  3. Once it is published, the Descriptron v2 paper as well. Its reference
+     will be added here and in CITATION.cff on publication.
+
+BibTeX:
+
+@software{vandam_2026_descriptron_v211,
+  author    = {Van Dam, Alex R.},
+  title     = {Descriptron: morphology-driven species descriptions, keys and
+               delimitation for dark taxa},
+  version   = {v2.1.1},
+  year      = {2026},
+  publisher = {Zenodo},
+  doi       = {10.5281/zenodo.22958109},
+  url       = {https://doi.org/10.5281/zenodo.22958109}
+}
+
+@article{vandam2026descriptron,
+  author  = {Van Dam, Alex R. and Štarhová Serbina, Liliya},
+  title   = {Descriptron: Artificial intelligence for automating taxonomic species
+             descriptions with a user-friendly software package},
+  journal = {Systematic Entomology},
+  volume  = {51},
+  number  = {1},
+  pages   = {e70005},
+  year    = {2026},
+  doi     = {10.1111/syen.70005}
+}
+
+The same information is in CITATION.cff, which GitHub shows as
+"Cite this repository".
+
+THIRD-PARTY FILES IN descriptron-sam3
+-------------------------------------
+tools/sam3_assets/bpe_simple_vocab_16e6.txt.gz - the byte-pair-encoding vocabulary of OpenAI CLIP
+(https://github.com/openai/CLIP), MIT licence, Copyright (c) 2021 OpenAI. SAM 3's text encoder needs it and the
+sam3 0.1.4 wheel on PyPI does not include it.
+
+SAM 3 itself (the `sam3` package, installed as a dependency) is Meta's, under the SAM License; its model
+checkpoints are gated and are downloaded by each user from https://huggingface.co/facebook/sam3 after requesting
+access. Neither the SAM 3 code nor its checkpoints are redistributed here.