streamlit-octostar-utils 0.6.5.dev2__tar.gz → 0.6.6__tar.gz

{streamlit_octostar_utils-0.6.5.dev2 → streamlit_octostar_utils-0.6.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: streamlit-octostar-utils
-Version: 0.6.5.dev2
+Version: 0.6.6
 Summary: 
 License: MIT
 License-File: LICENSE
@@ -24,6 +24,7 @@ Requires-Dist: iso639-lang (>=2.0.0,<3.0.0) ; extra == "nlp"
 Requires-Dist: nltk (>=3.8.0,<4.0.0) ; extra == "nlp"
 Requires-Dist: numpy (>=1.20.0)
 Requires-Dist: octostar-streamlit (>=0.1.25,<0.2.0)
+Requires-Dist: pillow (>=10.0.0)
 Requires-Dist: pottery (>=3.0.0,<4.0.0)
 Requires-Dist: presidio-analyzer (>=2.2.0,<3.0.0) ; extra == "nlp"
 Requires-Dist: psutil (>=5.9.0)

{streamlit_octostar_utils-0.6.5.dev2 → streamlit_octostar_utils-0.6.6}/pyproject.toml

@@ -5,7 +5,7 @@ include = '\.pyi?$'
 
 [tool.poetry]
 name = "streamlit-octostar-utils"
-version = "0.6.5-dev.2"
+version = "0.6.6"
 description = ""
 license = "MIT"
 authors = ["Octostar"]
@@ -43,6 +43,7 @@ pottery = "^3.0.0"
 slowapi = "^0.1.9"
 filetype = "^1.2.0"
 PyJWT = "^2.5.0"
+pillow = ">=10.0.0"
 #octostar-python-client = "^1.0.0"
 py3langid = { version = ">=0.2.0,<0.3.0", optional = true }
 spacy-download = { version = "==1.1.0", optional = true }

streamlit_octostar_utils-0.6.6/streamlit_octostar_utils/metadata/__init__.py

@@ -0,0 +1,17 @@
+from streamlit_octostar_utils.metadata.parent_meta import (
+    EXIF_PASSTHROUGH_FIELDS,
+    OS_PARENT_META_PREFIX,
+    embed_png_parent_meta,
+    ffmpeg_metadata_args,
+    resolve_with_aliases,
+    select_parent_meta_passthrough,
+)
+
+__all__ = [
+    "EXIF_PASSTHROUGH_FIELDS",
+    "OS_PARENT_META_PREFIX",
+    "embed_png_parent_meta",
+    "ffmpeg_metadata_args",
+    "resolve_with_aliases",
+    "select_parent_meta_passthrough",
+]

streamlit_octostar_utils-0.6.6/streamlit_octostar_utils/metadata/parent_meta.py

@@ -0,0 +1,152 @@
+"""Parent->child EXIF passthrough protocol.
+
+Producer routes that create new derived files (face crops, video keyframes,
+extracted audio) carry the parent's EXIF date/GPS tags through to downstream
+extract-metadata by embedding ``os_parent_meta_<canonical>`` keys into the
+child file's metadata container:
+
+* PNG: ``tEXt`` chunks via :func:`embed_png_parent_meta`.
+* Audio/video containers: ffmpeg ``-metadata`` flags via
+  :func:`ffmpeg_metadata_args`.
+
+The reader recovers them through :func:`resolve_with_aliases`, which is
+prefix-aware so it works whether exiftool surfaces the embedded key directly
+or with a format-specific group prefix (``PNG:``, ``ID3:``, ``RIFF:``,
+``pdf:docinfo:custom:``, ...).
+
+Keep producer and reader on the same version of this module: the suffix-match
+in :func:`resolve_with_aliases` is the only piece that knows what prefixes
+the various containers slap on, and a new container format means changes here
+that both sides must see.
+"""
+
+from io import BytesIO
+from typing import Any, Dict, List
+
+
+OS_PARENT_META_PREFIX = "os_parent_meta_"
+
+_PDF_CUSTOM_PREFIX = "pdf:docinfo:custom:"
+
+EXIF_PASSTHROUGH_FIELDS = (
+    "EXIF:DateTimeOriginal",
+    "Composite:SubSecDateTimeOriginal",
+    "Composite:DateTimeOriginal",
+    "EXIF:CreateDate",
+    "Composite:SubSecCreateDate",
+    "QuickTime:CreateDate",
+    "QuickTime:MediaCreateDate",
+    "QuickTime:TrackCreateDate",
+    "PDF:CreateDate",
+    "XMP:CreateDate",
+    "XMP:DateCreated",
+    "IPTC:DateCreated",
+    "Composite:DateTimeCreated",
+    "EXIF:DateTime",
+    "EXIF:ModifyDate",
+    "Composite:SubSecModifyDate",
+    "PDF:ModDate",
+    "XMP:ModifyDate",
+    "Composite:GPSLatitude",
+    "EXIF:GPSLatitude",
+    "XMP:GPSLatitude",
+    "Composite:GPSLatitudeRef",
+    "EXIF:GPSLatitudeRef",
+    "XMP:GPSLatitudeRef",
+    "Composite:GPSLongitude",
+    "EXIF:GPSLongitude",
+    "XMP:GPSLongitude",
+    "Composite:GPSLongitudeRef",
+    "EXIF:GPSLongitudeRef",
+    "XMP:GPSLongitudeRef",
+    "Composite:GPSPosition",
+)
+
+
+def resolve_with_aliases(metadata: Dict[str, Any], key: str) -> Any:
+    """Return ``metadata[key]`` respecting ``os_parent_meta_<key>`` passthrough.
+
+    Lookup order:
+
+    1. ``os_parent_meta_<key>`` -- written directly by producer routes that
+       can place a key into the child container under an exact name.
+    2. ``pdf:docinfo:custom:os_parent_meta_<key>`` -- PDF custom property
+       container (mirrors the ``os_ContextBefore`` pattern used elsewhere).
+    3. Any key ending with ``:os_parent_meta_<key>`` -- catches the
+       format-specific group prefix that exiftool prepends when reading
+       container-native tags (``PNG:`` for PNG tEXt, ``ID3:`` for MP3 TXXX,
+       ``RIFF:`` for WAV INFO, etc.).
+    4. ``<key>`` -- canonical fallback (the value as it would appear if the
+       child carried real EXIF natively).
+    """
+    aliased = f"{OS_PARENT_META_PREFIX}{key}"
+    value = metadata.get(aliased)
+    if value is not None:
+        return value
+    value = metadata.get(f"{_PDF_CUSTOM_PREFIX}{aliased}")
+    if value is not None:
+        return value
+    suffix = f":{aliased}"
+    for k, v in metadata.items():
+        if v is not None and k.endswith(suffix):
+            return v
+    return metadata.get(key)
+
+
+def select_parent_meta_passthrough(metadata: Any) -> Dict[str, str]:
+    """Pick EXIF date/GPS tags worth embedding into derived children.
+
+    Input is a parent's raw exiftool metadata dict (as stored on the parent
+    entity's ``extract:metadata`` annotation). Output is keyed by
+    ``os_parent_meta_<canonical>`` with stringified values, ready to feed
+    into a container-specific embedder.
+
+    Uses :func:`resolve_with_aliases` so a parent that itself inherited tags
+    from *its* parent (e.g. a keyframe of an inherited-meta video) still
+    propagates the original values forward.
+    """
+    if not isinstance(metadata, dict):
+        return {}
+    out: Dict[str, str] = {}
+    for key in EXIF_PASSTHROUGH_FIELDS:
+        value = resolve_with_aliases(metadata, key)
+        if value is None:
+            continue
+        if isinstance(value, (list, tuple)):
+            value = ",".join(str(v) for v in value)
+        out[f"{OS_PARENT_META_PREFIX}{key}"] = str(value)
+    return out
+
+
+def embed_png_parent_meta(png_bytes: bytes, parent_meta: Dict[str, str]) -> bytes:
+    """Add ``parent_meta`` keys as PNG ``tEXt`` chunks. No-op if empty."""
+    if not parent_meta:
+        return png_bytes
+    from PIL import Image
+    from PIL.PngImagePlugin import PngInfo
+
+    img = Image.open(BytesIO(png_bytes))
+    info = PngInfo()
+    for k, v in (img.info or {}).items():
+        if isinstance(v, str):
+            info.add_text(k, v)
+    for k, v in parent_meta.items():
+        info.add_text(k, v)
+    out = BytesIO()
+    img.save(out, format="PNG", pnginfo=info)
+    return out.getvalue()
+
+
+def ffmpeg_metadata_args(parent_meta: Dict[str, str]) -> List[str]:
+    """Return ``-metadata key=value`` ffmpeg flag pairs for ``parent_meta``.
+
+    ffmpeg writes these into whatever metadata container the output format
+    supports (ID3v2 TXXX for MP3, INFO chunk for WAV, custom atoms for
+    MP4/M4A, vorbis comments for OGG). exiftool surfaces them with the
+    container's group prefix (``ID3:``, ``RIFF:``, ...), which
+    :func:`resolve_with_aliases` handles via its suffix match.
+    """
+    args: List[str] = []
+    for k, v in (parent_meta or {}).items():
+        args.extend(["-metadata", f"{k}={v}"])
+    return args