robotensor-attune 0.2.1__py3-none-any.whl

robotensor_attune-0.2.1.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 2026 Robotensor contributors
+
+   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.

vector_runtime/__init__.py

@@ -0,0 +1,11 @@
+"""The validator's runtime for `vector_v1.1` weights.
+
+A submission is one `model.safetensors`: no code, no pickle. `check` holds its header to the
+architecture's tensors (standard library only), `net` builds the network and loads it, and
+`policy.VectorPolicy` serves it over `vector-protocol`.
+"""
+
+__version__ = "0.3.0"
+
+ARCHITECTURE = "vector_v1.1"
+WEIGHTS_FILENAME = "model.safetensors"

vector_runtime/check.py

@@ -0,0 +1,163 @@
+"""Is this `model.safetensors` exactly the architecture's tensors? Decided from the header.
+
+    vector-runtime check --weights DIR_OR_FILE [--max-bytes N] [--no-hash]
+
+The standard library only: nothing in the file is read beyond its JSON header. It passes when it
+is a regular file of at most `max_file_bytes` whose tensors tile the data section exactly and are
+exactly the architecture's (`<architecture>.json`: every key, shape and dtype). The report carries
+`weights_sha256`, the identity a result is recorded under.
+"""
+
+from __future__ import annotations
+
+import hashlib
+import json
+import os
+import stat
+from dataclasses import asdict, dataclass, field
+from pathlib import Path
+from typing import Any
+
+from . import ARCHITECTURE, WEIGHTS_FILENAME
+from .header import DEFAULT_MAX_HEADER_BYTES, HeaderError, read_header
+
+#: The largest file checked unless the caller says otherwise. A real one is ~3.1 GB.
+DEFAULT_MAX_FILE_BYTES = 8 * 1024**3
+#: How many missing, unexpected or mismatched tensors an error lists by name.
+LISTED = 20
+_CHUNK = 8 * 1024 * 1024
+
+
+@dataclass
+class CheckReport:
+    path: str
+    architecture: str
+    errors: list[str] = field(default_factory=list)
+    warnings: list[str] = field(default_factory=list)
+    file_bytes: int | None = None
+    header_bytes: int | None = None
+    tensor_count: int | None = None
+    param_count: int | None = None
+    metadata: dict[str, str] | None = None
+    weights_sha256: str | None = None
+    other_files: list[str] = field(default_factory=list)
+
+    @property
+    def ok(self) -> bool:
+        return not self.errors
+
+    def to_dict(self) -> dict[str, Any]:
+        return {"ok": self.ok, **asdict(self)}
+
+
+def weights_file(path: str | os.PathLike[str]) -> Path:
+    """`path` itself, or `path/model.safetensors` when it is a directory."""
+    path = Path(path)
+    return path / WEIGHTS_FILENAME if path.is_dir() else path
+
+
+def check(
+    weights: str | os.PathLike[str],
+    *,
+    architecture: str = ARCHITECTURE,
+    max_file_bytes: int = DEFAULT_MAX_FILE_BYTES,
+    max_header_bytes: int = DEFAULT_MAX_HEADER_BYTES,
+    compute_sha256: bool = True,
+) -> CheckReport:
+    """Check the weights at `weights` (a file, or a directory holding `model.safetensors`)."""
+    path = weights_file(weights)
+    report = CheckReport(path=str(path), architecture=architecture)
+    if Path(weights).is_dir():
+        report.other_files = sorted(
+            entry.name for entry in Path(weights).iterdir() if entry.name != WEIGHTS_FILENAME
+        )
+        if report.other_files:
+            report.warnings.append(
+                f"the directory holds more than {WEIGHTS_FILENAME}: {', '.join(report.other_files)}"
+            )
+    expected = tensors(architecture)
+
+    try:
+        info = path.stat()  # follows a symbolic link, as a Hugging Face snapshot holds them
+    except OSError as exc:
+        report.errors.append(f"cannot read {path}: {exc.strerror or exc}")
+        return report
+    if not stat.S_ISREG(info.st_mode):
+        report.errors.append(f"{path} is not a regular file")
+        return report
+    report.file_bytes = info.st_size
+    if info.st_size > max_file_bytes:
+        report.errors.append(f"the file is {info.st_size} bytes; at most {max_file_bytes}")
+        return report
+
+    try:
+        with path.open("rb") as handle:
+            header = read_header(handle, info.st_size, max_header_bytes=max_header_bytes)
+    except HeaderError as exc:
+        report.errors.append(f"not a valid safetensors file: {exc}")
+    except OSError as exc:
+        report.errors.append(f"cannot read {path}: {exc.strerror or exc}")
+        return report
+    else:
+        report.header_bytes = header.header_bytes
+        report.tensor_count = len(header.tensors)
+        report.param_count = header.param_count
+        report.metadata = header.metadata
+        report.errors.extend(compare(header.tensors, expected))
+
+    if compute_sha256:
+        try:
+            report.weights_sha256 = sha256_file(path)
+        except OSError as exc:
+            report.errors.append(f"cannot hash {path}: {exc.strerror or exc}")
+    return report
+
+
+#: The name `vector_orchestrator.duel.weights_runtime` calls it by.
+check_weights = check
+
+
+def manifest(architecture: str = ARCHITECTURE) -> Path:
+    return Path(__file__).resolve().parent / f"{architecture}.json"
+
+
+def tensors(architecture: str = ARCHITECTURE) -> dict[str, dict[str, Any]]:
+    """`{key: {"shape": [...], "dtype": "F32"}}`: every tensor the architecture's file holds."""
+    return json.loads(manifest(architecture).read_text(encoding="utf-8"))
+
+
+def compare(tensors: dict[str, Any], expected: dict[str, dict[str, Any]]) -> list[str]:
+    """Every way the header's tensors differ from the architecture's, as error messages."""
+    errors: list[str] = []
+    missing = sorted(set(expected) - set(tensors))
+    unexpected = sorted(set(tensors) - set(expected))
+    for key in missing[:LISTED]:
+        errors.append(f"tensor {key} is missing")
+    for key in unexpected[:LISTED]:
+        errors.append(f"tensor {key} is not part of the architecture")
+    mismatched = 0
+    for key in sorted(set(tensors) & set(expected)):
+        got, want = tensors[key], expected[key]
+        problems = []
+        if list(got.shape) != list(want["shape"]):
+            problems.append(f"shape {list(got.shape)} != {list(want['shape'])}")
+        if got.dtype != want["dtype"]:
+            problems.append(f"dtype {got.dtype} != {want['dtype']}")
+        if problems:
+            mismatched += 1
+            if mismatched <= LISTED:
+                errors.append(f"tensor {key}: {', '.join(problems)}")
+    if len(missing) > LISTED or len(unexpected) > LISTED or mismatched > LISTED:
+        errors.append(
+            f"in all: {len(missing)} missing, {len(unexpected)} unexpected and {mismatched} "
+            f"mismatched of the {len(expected)} tensors the architecture has"
+        )
+    return errors
+
+
+def sha256_file(path: str | os.PathLike[str]) -> str:
+    digest = hashlib.sha256()
+    with open(path, "rb") as handle:
+        while chunk := handle.read(_CHUNK):
+            digest.update(chunk)
+    return digest.hexdigest()

vector_runtime/header.py

@@ -0,0 +1,191 @@
+"""A safetensors file's header, parsed by hand: the standard library only, no tensor data read.
+
+The format: an 8-byte little-endian unsigned length N, then N bytes of UTF-8 JSON, then the data.
+The JSON maps each tensor name to `{"dtype", "shape", "data_offsets": [begin, end]}`, offsets
+relative to the start of the data, plus an optional `"__metadata__"` of string to string.
+
+The file is untrusted, so the parse is strict where a lenient one would let two readers disagree
+about what the file holds: a duplicated key, a NaN, an unknown dtype, a byte range that does not
+match its shape, a gap or an overlap between tensors, or data past the last tensor are all
+refused. What is accepted is exactly what `safetensors` itself would load, tensor for tensor.
+"""
+
+from __future__ import annotations
+
+import json
+import struct
+from dataclasses import dataclass
+from typing import Any, BinaryIO
+
+#: Bytes per element of every dtype safetensors defines.
+DTYPE_SIZES = {
+    "BOOL": 1,
+    "U8": 1,
+    "I8": 1,
+    "F8_E5M2": 1,
+    "F8_E4M3": 1,
+    "I16": 2,
+    "U16": 2,
+    "F16": 2,
+    "BF16": 2,
+    "I32": 4,
+    "U32": 4,
+    "F32": 4,
+    "I64": 8,
+    "U64": 8,
+    "F64": 8,
+}
+METADATA_KEY = "__metadata__"
+LENGTH_BYTES = 8
+#: The largest header read unless the caller says otherwise. The real one is ~0.1 MB.
+DEFAULT_MAX_HEADER_BYTES = 16 * 1024 * 1024
+#: The most dimensions a tensor may declare.
+MAX_RANK = 16
+
+
+class HeaderError(ValueError):
+    """The bytes are not a well-formed safetensors header."""
+
+
+@dataclass(frozen=True)
+class TensorInfo:
+    dtype: str
+    shape: tuple[int, ...]
+    begin: int
+    end: int
+
+    @property
+    def numel(self) -> int:
+        n = 1
+        for dim in self.shape:
+            n *= dim
+        return n
+
+
+@dataclass(frozen=True)
+class Header:
+    """A parsed header: its JSON length, the tensors, the metadata and the data section's size."""
+
+    header_bytes: int
+    tensors: dict[str, TensorInfo]
+    metadata: dict[str, str] | None = None
+    data_bytes: int = 0
+
+    @property
+    def total_bytes(self) -> int:
+        return LENGTH_BYTES + self.header_bytes + self.data_bytes
+
+    @property
+    def param_count(self) -> int:
+        return sum(t.numel for t in self.tensors.values())
+
+
+def read_header(
+    handle: BinaryIO, file_size: int, *, max_header_bytes: int = DEFAULT_MAX_HEADER_BYTES
+) -> Header:
+    """The header of the safetensors file open as `handle` (at offset 0), `file_size` bytes long.
+
+    Raises `HeaderError` for anything that is not a well-formed file of exactly that size.
+    """
+    prefix = handle.read(LENGTH_BYTES)
+    if len(prefix) != LENGTH_BYTES:
+        raise HeaderError(f"the file is {file_size} bytes, too short for a safetensors header")
+    (length,) = struct.unpack("<Q", prefix)
+    if length < 2:
+        raise HeaderError(f"the header length is {length} bytes, too short for a JSON object")
+    if length > max_header_bytes:
+        raise HeaderError(f"the header is {length} bytes; at most {max_header_bytes} are read")
+    if length > file_size - LENGTH_BYTES:
+        raise HeaderError(
+            f"the header claims {length} bytes but the file has {file_size - LENGTH_BYTES} after "
+            "the length"
+        )
+    raw = handle.read(length)
+    if len(raw) != length:
+        raise HeaderError(f"the header was cut short: {len(raw)} of {length} bytes")
+    return parse_header(raw, file_size)
+
+
+def parse_header(raw: bytes, file_size: int) -> Header:
+    """The header whose JSON bytes are `raw`, in a file of `file_size` bytes."""
+    try:
+        text = raw.decode("utf-8")
+    except UnicodeDecodeError as exc:
+        raise HeaderError(f"the header is not UTF-8: {exc}") from None
+    if not text.startswith("{"):
+        raise HeaderError("the header does not start with a JSON object")
+    try:
+        document = json.loads(
+            text, object_pairs_hook=_unique_object, parse_constant=_refuse_constant
+        )
+    except HeaderError:
+        raise
+    except (ValueError, RecursionError) as exc:
+        raise HeaderError(f"the header is not valid JSON: {exc}") from None
+    if not isinstance(document, dict):
+        raise HeaderError("the header is not a JSON object")
+
+    metadata = document.pop(METADATA_KEY, None)
+    if metadata is not None:
+        if not isinstance(metadata, dict) or not all(
+            isinstance(k, str) and isinstance(v, str) for k, v in metadata.items()
+        ):
+            raise HeaderError(f"{METADATA_KEY} must map strings to strings")
+
+    tensors = {name: _tensor(name, entry) for name, entry in document.items()}
+    data_bytes = file_size - LENGTH_BYTES - len(raw)
+    _check_layout(tensors, data_bytes)
+    return Header(header_bytes=len(raw), tensors=tensors, metadata=metadata, data_bytes=data_bytes)
+
+
+def _tensor(name: str, entry: Any) -> TensorInfo:
+    if not name:
+        raise HeaderError("a tensor has an empty name")
+    if not isinstance(entry, dict) or set(entry) != {"dtype", "shape", "data_offsets"}:
+        raise HeaderError(f"{name}: expected exactly dtype, shape and data_offsets")
+    dtype, shape, offsets = entry["dtype"], entry["shape"], entry["data_offsets"]
+    if dtype not in DTYPE_SIZES:
+        raise HeaderError(f"{name}: unknown dtype {dtype!r}")
+    if not isinstance(shape, list) or len(shape) > MAX_RANK or not all(_count(d) for d in shape):
+        raise HeaderError(f"{name}: shape must be a list of at most {MAX_RANK} non-negative ints")
+    if not isinstance(offsets, list) or len(offsets) != 2 or not all(_count(o) for o in offsets):
+        raise HeaderError(f"{name}: data_offsets must be two non-negative ints")
+    begin, end = offsets
+    info = TensorInfo(dtype=dtype, shape=tuple(shape), begin=begin, end=end)
+    if end < begin or end - begin != info.numel * DTYPE_SIZES[dtype]:
+        raise HeaderError(
+            f"{name}: data_offsets {offsets} hold {end - begin} bytes; a {dtype} tensor of shape "
+            f"{shape} is {info.numel * DTYPE_SIZES[dtype]}"
+        )
+    return info
+
+
+def _check_layout(tensors: dict[str, TensorInfo], data_bytes: int) -> None:
+    """The tensors tile the data section exactly: no gap, no overlap, nothing after the last."""
+    cursor = 0
+    for name, info in sorted(tensors.items(), key=lambda item: (item[1].begin, item[1].end)):
+        if info.begin != cursor:
+            what = "overlaps the tensor before it" if info.begin < cursor else "leaves a gap"
+            raise HeaderError(f"{name}: starts at byte {info.begin} and {what} (expected {cursor})")
+        cursor = info.end
+    if cursor != data_bytes:
+        raise HeaderError(
+            f"the tensors hold {cursor} bytes of data but the file has {data_bytes} after the header"
+        )
+
+
+def _count(value: Any) -> bool:
+    return type(value) is int and value >= 0
+
+
+def _unique_object(pairs: list[tuple[str, Any]]) -> dict[str, Any]:
+    out: dict[str, Any] = {}
+    for key, value in pairs:
+        if key in out:
+            raise HeaderError(f"the header has the key {key!r} twice")
+        out[key] = value
+    return out
+
+
+def _refuse_constant(name: str) -> Any:
+    raise HeaderError(f"the header holds {name}, which is not JSON")