classifyre-cli 0.4.136__tar.gz → 0.4.138__tar.gz

classifyre_cli-0.4.138/.turbo/turbo-build.log

@@ -0,0 +1,3 @@
+$ uv sync
+Resolved 322 packages in 301ms
+Checked 51 packages in 1ms

{classifyre_cli-0.4.136 → classifyre_cli-0.4.138}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: classifyre-cli
-Version: 0.4.136
+Version: 0.4.138
 Summary: Classifyre CLI — scan and classify unstructured data sources
 License: MIT
 Keywords: data,ingestion,metadata,pii,secrets,unstructured
@@ -43,14 +43,11 @@ docx|xlsx|pptx` and feeds the result through the normal OOXML extractors. Withou
 it those files scan as *"no content available"* — a silent coverage gap, not an
 error.
 
-- **Kubernetes**: installed in the CLI image (`libreoffice-*-nogui`), so this
-  works out of the box.
-- **Desktop and local dev**: install it yourself — `brew install --cask libreoffice`
+- **Kubernetes and the all-in-one Docker image**: installed in the image
+  (`libreoffice-*-nogui`), so this works out of the box.
+- **Local dev**: install it yourself — `brew install --cask libreoffice`
   (macOS), `apt install libreoffice-writer libreoffice-calc libreoffice-impress`
-  (Debian/Ubuntu), or libreoffice.org on Windows. The desktop app deliberately
-  does not bundle LibreOffice: it is ~550 MB even stripped, and slimming the
-  upstream bundle breaks its code signature, which Apple Silicon punishes by
-  killing the process on launch.
+  (Debian/Ubuntu), or libreoffice.org on Windows.
 - **Non-standard install**: set `CLASSIFYRE_SOFFICE_PATH` to the binary. The CLI
   checks that first, then `PATH`, then the standard macOS, Windows, and
   Linux-tarball install locations.

{classifyre_cli-0.4.136 → classifyre_cli-0.4.138}/README.md

@@ -26,14 +26,11 @@ docx|xlsx|pptx` and feeds the result through the normal OOXML extractors. Withou
 it those files scan as *"no content available"* — a silent coverage gap, not an
 error.
 
-- **Kubernetes**: installed in the CLI image (`libreoffice-*-nogui`), so this
-  works out of the box.
-- **Desktop and local dev**: install it yourself — `brew install --cask libreoffice`
+- **Kubernetes and the all-in-one Docker image**: installed in the image
+  (`libreoffice-*-nogui`), so this works out of the box.
+- **Local dev**: install it yourself — `brew install --cask libreoffice`
   (macOS), `apt install libreoffice-writer libreoffice-calc libreoffice-impress`
-  (Debian/Ubuntu), or libreoffice.org on Windows. The desktop app deliberately
-  does not bundle LibreOffice: it is ~550 MB even stripped, and slimming the
-  upstream bundle breaks its code signature, which Apple Silicon punishes by
-  killing the process on launch.
+  (Debian/Ubuntu), or libreoffice.org on Windows.
 - **Non-standard install**: set `CLASSIFYRE_SOFFICE_PATH` to the binary. The CLI
   checks that first, then `PATH`, then the standard macOS, Windows, and
   Linux-tarball install locations.

{classifyre_cli-0.4.136 → classifyre_cli-0.4.138}/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@classifyre/cli",
-  "version": "0.4.136",
+  "version": "0.4.138",
   "private": true,
   "scripts": {
     "build": "uv sync",

{classifyre_cli-0.4.136 → classifyre_cli-0.4.138}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "classifyre-cli"
-version = "0.4.136"
+version = "0.4.138"
 description = "Classifyre CLI — scan and classify unstructured data sources"
 readme = "README.md"
 requires-python = ">=3.12"
@@ -474,10 +474,10 @@ package = true
 # distribution declared by our video and OCR groups. Installing both would
 # overwrite the same cv2 namespace and add GUI libraries to CPU-only containers.
 exclude-dependencies = ["opencv-python"]
-# Every platform we ship (desktop bundles + CI) must resolve to versions with
+# Every platform we ship must resolve to versions with
 # usable wheels. Without this, uv happily locks versions that are missing a
 # platform — e.g. deltalake 1.6.1 published no win_amd64 wheel, which broke the
-# Windows desktop venv bake ("can't be installed because it doesn't have a
+# Windows venv bake ("can't be installed because it doesn't have a
 # source distribution or wheel for the current platform").
 required-environments = [
     "sys_platform == 'win32' and platform_machine == 'AMD64'",

{classifyre_cli-0.4.136 → classifyre_cli-0.4.138}/src/augmentation/session.py

@@ -223,7 +223,7 @@ class AugmentationSession:
             if not path.is_dir():
                 raise ChildProcessError(
                     f"Local folder {entry.get('name')!r} points at {path}, which is not "
-                    "a directory. Local folders are only available in the desktop application."
+                    "a directory. The folder must be mounted where the scan runs."
                 )
 
     async def _slot(self, index: int) -> _ChildSlot:

{classifyre_cli-0.4.136 → classifyre_cli-0.4.138}/src/models/generated_input.py

@@ -1031,7 +1031,7 @@ class LocalFolderRequired(BaseModel):
     )
     path: str = Field(
         ...,
-        description="Absolute path of the folder to scan, as seen by the machine that runs the scan. On desktop that is this computer (for example /Users/jane/Documents). In Kubernetes it is the scan pod, so the folder must be mounted into CLI jobs first via the chart's `api.localFolders` — then use the `mountPath` you gave it there (for example /mnt/corpora/invoices).",
+        description="Absolute path of the folder to scan, as seen by the machine that runs the scan — not as your browser sees it. With the all-in-one Docker image, bind-mount the directory and use the container-side path (`-v ~/invoices:/data/invoices:ro` → /data/invoices). In Kubernetes the scan runs in a pod, so the folder must be mounted into CLI jobs first via the chart's `api.localFolders` — then use the `mountPath` you gave it there (for example /mnt/corpora/invoices).",
     )
 
 
@@ -1063,7 +1063,7 @@ class SandboxOptional(BaseModel):
 
 class LocalFolderMasked(BaseModel):
     """
-    A folder scan reads the filesystem it runs on with that process's own permissions — the desktop app's, or the CLI job pod's. No credentials are required.
+    A folder scan reads the filesystem it runs on with that process's own permissions — the container's, or the CLI job pod's. No credentials are required.
     """
 
     model_config = ConfigDict(
@@ -4654,7 +4654,7 @@ class NotebookPackage(BaseModel):
 
 class NotebookLocalFolder(BaseModel):
     """
-    A folder the notebook can read on the machine that runs the scan: this computer on desktop, or a path mounted into CLI jobs by the chart's `api.localFolders` in Kubernetes.
+    A folder the notebook can read on the machine that runs the scan: a bind-mounted path in the all-in-one Docker image, or one mounted into CLI jobs by the chart's `api.localFolders` in Kubernetes.
     """
 
     model_config = ConfigDict(
@@ -4667,7 +4667,7 @@ class NotebookLocalFolder(BaseModel):
     )
     path: str = Field(
         ...,
-        description='Absolute path to the folder, as seen by the machine that runs the scan — for example /Users/me/dumps on desktop, or /mnt/corpora/dumps for a folder mounted by the chart.',
+        description='Absolute path to the folder, as seen by the machine that runs the scan — for example /data/dumps for a directory bind-mounted into the container, or /mnt/corpora/dumps for a folder mounted by the chart.',
         max_length=4096,
         min_length=1,
     )
@@ -4777,7 +4777,7 @@ class AugmentationConfig(BaseModel):
     )
     local_folders: list[NotebookLocalFolder] | None = Field(
         [],
-        description='Desktop only. Folders on this machine the notebook reads with ctx.folder("name"). Not available in Kubernetes deployments, where files are uploaded to the source instead.',
+        description='Folders the notebook reads with ctx.folder("name"). Not available in Kubernetes deployments, where files are uploaded to the source instead.',
         max_length=10,
         validate_default=True,
     )
@@ -5709,7 +5709,7 @@ class CustomOptional(BaseModel):
     )
     local_folders: list[NotebookLocalFolder] | None = Field(
         [],
-        description='Desktop only. Folders on this machine the notebook reads with ctx.folder("name"). Not available in Kubernetes deployments, where files are uploaded to the source instead. This is a convenience, not a sandbox: the notebook process runs as you and can open any path you can.',
+        description='Folders the notebook reads with ctx.folder("name"). Not available in Kubernetes deployments, where files are uploaded to the source instead. This is a convenience, not a sandbox: the notebook process runs as you and can open any path you can.',
         max_length=10,
         validate_default=True,
     )

{classifyre_cli-0.4.136 → classifyre_cli-0.4.138}/src/notebook/files.py

@@ -59,7 +59,7 @@ class ParsedContent:
 
 
 def local_folders(recipe: dict[str, Any]) -> dict[str, str]:
-    """The desktop-only folders a source declared, as ``{name: path}``.
+    """The folders a source declared, as ``{name: path}``.
 
     Read straight from the recipe rather than handed over separately: unlike an
     uploaded file there is nothing to fetch, only a path the notebook may open.

{classifyre_cli-0.4.136 → classifyre_cli-0.4.138}/src/notebook/sdk.py

@@ -16,8 +16,8 @@ the only place the distinction can come from.
 .eml or a Parquet file is not connector logic, every other source in the system
 already delegates it to the same parser, and a notebook that had to reimplement
 it would get it wrong. It arrives with the files it operates on: ``ctx.files``
-for what was uploaded to the source, ``ctx.folder(...)`` for a folder on a
-desktop machine.
+for what was uploaded to the source, ``ctx.folder(...)`` for a folder mounted
+into the machine that runs the scan.
 """
 
 from __future__ import annotations
@@ -331,10 +331,10 @@ class Context:
     def folders(self) -> dict[str, Path]:
         """Folders this source was configured with, by name.
 
-        The paths are resolved on whatever machine runs the scan: this computer
-        on desktop, an ephemeral CLI job pod in Kubernetes. In a cluster the
-        folder has to be mounted into those pods first (the chart's
-        ``api.localFolders``); an unmounted path simply will not exist.
+        The paths are resolved on whatever machine runs the scan: inside the
+        all-in-one container, or an ephemeral CLI job pod in Kubernetes. Either
+        way the folder has to be mounted there first (a bind mount, or the
+        chart's ``api.localFolders``); an unmounted path simply will not exist.
         """
         return dict(self._folders)
 

{classifyre_cli-0.4.136 → classifyre_cli-0.4.138}/src/outputs/rest.py

@@ -20,7 +20,7 @@ logger = logging.getLogger(__name__)
 # The API has no user authentication, so the ingest endpoints are restricted to
 # callers holding this shared secret; the API injects it into every CLI job's
 # environment (Kubernetes job env, or inherited from the parent process on
-# desktop). Absent locally, where the API leaves the restriction disabled.
+# all-in-one image). Absent locally, where the API leaves the restriction disabled.
 INTERNAL_KEY_HEADER = "X-Classifyre-Internal-Key"
 
 

{classifyre_cli-0.4.136 → classifyre_cli-0.4.138}/src/pipeline/worker_pool.py

@@ -45,7 +45,7 @@ def _init_worker_threads(threads_per_worker: int) -> None:
     threads = max(1, threads_per_worker)
     for var in ("OMP_NUM_THREADS", "MKL_NUM_THREADS", "OPENBLAS_NUM_THREADS"):
         os.environ.setdefault(var, str(threads))
-    # Detector work is background work: on a shared host (desktop) the pool
+    # Detector work is background work: on a shared host the pool
     # must yield to the UI, so deprioritize workers where the OS allows it.
     if hasattr(os, "nice"):
         try:
@@ -235,7 +235,7 @@ class DetectorWorkerPool:
 
         ctx = multiprocessing.get_context(mp_start_method)
         # CLASSIFYRE_WORKER_THREADS caps BLAS/torch threads per worker; the
-        # desktop app sets it so workers*threads stays below the core count
+        # the all-in-one image sets it so workers*threads stays below the core count
         # instead of saturating the machine. Unset (K8s/dev) keeps the
         # divide-the-cores behavior.
         threads_env = os.environ.get("CLASSIFYRE_WORKER_THREADS", "")

{classifyre_cli-0.4.136 → classifyre_cli-0.4.138}/src/sources/custom/source.py

@@ -233,7 +233,7 @@ class CustomSource(BaseSource):
             if not path.is_dir():
                 raise CustomSourceError(
                     f"Local folder {folder.name!r} points at {path}, which is not a directory. "
-                    "Local folders are only available in the desktop application."
+                    "The folder must be mounted where the scan runs."
                 )
 
     def _materialize_uploaded_files(self) -> Path | None:

{classifyre_cli-0.4.136 → classifyre_cli-0.4.138}/src/sources/git/source.py

@@ -10,7 +10,7 @@ Two properties drive the design:
 directory with its own ``HOME`` and its own git configuration, and deletes all of
 it when the run ends — normally, on abort, and on failure. There is no clone
 cache and no credential cache, so one job can never see another job's objects or
-another job's secrets. In Kubernetes the pod dies anyway; on the desktop, where
+another job's secrets. In Kubernetes the pod dies anyway; in a long-lived container, where
 the process is long-lived and scans many sources, that guarantee has to be made
 by this module.
 
@@ -573,7 +573,7 @@ class GitSource(ObjectStorageSourceBase):
     def _git_binary() -> str:
         """The git executable to use.
 
-        ``CLASSIFYRE_GIT_BINARY`` lets a packaged desktop build point at the git
+        ``CLASSIFYRE_GIT_BINARY`` lets a packaged build point at the git
         it ships rather than whatever the user happens to have installed — an
         operational concern, so it is an environment variable rather than a
         field on the source.

{classifyre_cli-0.4.136 → classifyre_cli-0.4.138}/src/sources/local_folder/source.py

@@ -16,7 +16,7 @@ logger = logging.getLogger(__name__)
 
 
 class LocalFolderSource(ObjectStorageSourceBase):
-    """Scans a folder on the local filesystem (desktop application deployments).
+    """Scans a folder on the filesystem the scan process runs on.
 
     Reuses the object-storage pipeline: each file becomes an ObjectRef whose key
     is the path relative to the scanned root, so sampling, MIME resolution, text

{classifyre_cli-0.4.136 → classifyre_cli-0.4.138}/src/utils/legacy_office.py

@@ -8,7 +8,7 @@ is fed back through the existing docx/xlsx/pptx extraction paths:
 
 The CLI container image ships LibreOffice (the ``libreoffice-*-nogui`` Debian
 packages, which install ``/usr/bin/soffice``), so conversion works out of the box
-under Kubernetes. The desktop app deliberately does NOT bundle it — ~550 MB even
+under Kubernetes, and in the all-in-one Docker image. A packaged distribution may deliberately NOT bundle it — ~550 MB even
 stripped, and slimming the upstream bundle breaks its code signature — so there
 it is a system dependency the user installs.
 
@@ -55,7 +55,7 @@ LEGACY_OFFICE_MIME_TYPES = frozenset(_CONVERSION_TARGETS)
 
 _SOFFICE_TIMEOUT_SECONDS = 120
 
-# How to actually get LibreOffice, per platform. The desktop app does not bundle
+# How to actually get LibreOffice, per platform. A distribution that does not bundle
 # it, so this message is the entire remediation path a user gets — a generic
 # "install LibreOffice" would leave them guessing which package to install.
 _INSTALL_HINTS: dict[str, str] = {
@@ -85,14 +85,14 @@ def soffice_missing_error() -> str:
     )
 
 
-# Explicit override, checked before anything else. The desktop app rebuilds a
+# Explicit override, checked before anything else. A packaged distribution rebuilds a
 # minimal PATH for the processes it spawns, so an installation it discovered
 # itself is handed down through this variable rather than through PATH.
 SOFFICE_PATH_ENV = "CLASSIFYRE_SOFFICE_PATH"
 
 # Install locations to check when PATH comes up empty, keyed by sys.platform.
 # macOS and Windows always need these: their installers drop the binary inside an
-# app bundle / Program Files without touching PATH, and the packaged desktop app
+# app bundle / Program Files without touching PATH, and a packaged distribution
 # rebuilds a minimal PATH that cannot see either. Linux only needs them for an
 # upstream tarball install — the distro packages (including the
 # libreoffice-*-nogui ones in the CLI image) all land on /usr/bin/soffice.
@@ -119,7 +119,7 @@ def find_soffice() -> str | None:
 
     Resolution order, most explicit first:
 
-    1. ``CLASSIFYRE_SOFFICE_PATH`` — an operator/desktop-supplied absolute path.
+    1. ``CLASSIFYRE_SOFFICE_PATH`` — an operator-supplied absolute path.
     2. ``PATH`` — the CLI container, Linux hosts, and dev shells.
     3. Platform install locations — macOS/Windows only (see the table above).
 

{classifyre_cli-0.4.136 → classifyre_cli-0.4.138}/src/utils/resources.py

@@ -63,7 +63,7 @@ def get_effective_memory_mb() -> int:
         pass
 
     # macOS (and other POSIX without /proc): total physical memory via sysconf.
-    # Without this the desktop app fell through to the 4096 MB guess, which
+    # Without this a single-host install fell through to the 4096 MB guess, which
     # oversized the pool on small laptops and undersized it on big ones.
     try:
         pages = os.sysconf("SC_PHYS_PAGES")

{classifyre_cli-0.4.136 → classifyre_cli-0.4.138}/src/utils/torch_runtime.py

@@ -5,9 +5,9 @@ generates C++ and shells out to a compiler *at scan time*. That is a poor fit
 for how this CLI is deployed, in two independent ways:
 
 1. It needs a working C++ toolchain on the machine running the scan. Neither
-   the desktop bundle nor a slim container image promises one.
+   a packaged bundle nor a slim container image promises one.
 2. Inductor builds the compiler command line without quoting its paths, so any
-   path containing a space breaks it. The desktop venv lives under
+   path containing a space breaks it. A venv living under
    ``~/Library/Application Support/…``, which produces exactly that:
 
        clang++: error: no such file or directory: