mostlyright-data 0.9.0__tar.gz → 0.9.2__tar.gz

mostlyright_data-0.9.2/PKG-INFO

@@ -0,0 +1,166 @@
+Metadata-Version: 2.4
+Name: mostlyright-data
+Version: 0.9.2
+Summary: Mostly Right CLI for reproducible, reviewed Parquet datasets
+Project-URL: Homepage, https://mostlyright.md/
+Project-URL: Documentation, https://mostlyright.md/docs/guides/cli/
+Requires-Python: <3.14,>=3.13
+Provides-Extra: build
+Requires-Dist: hatchling==1.28.0; extra == 'build'
+Requires-Dist: uv-build==0.9.26; extra == 'build'
+Provides-Extra: dev
+Requires-Dist: cmarkgfm>=2025.10.22; extra == 'dev'
+Requires-Dist: croniter==6.2.4; extra == 'dev'
+Requires-Dist: cryptography<51,>=50.0.0; extra == 'dev'
+Requires-Dist: et-xmlfile==2.0.0; extra == 'dev'
+Requires-Dist: idna<4,>=3.10; extra == 'dev'
+Requires-Dist: jsonschema[format]<5,>=4.23; extra == 'dev'
+Requires-Dist: numpy==2.5.1; extra == 'dev'
+Requires-Dist: openpyxl==3.1.5; extra == 'dev'
+Requires-Dist: packaging<27,>=24; extra == 'dev'
+Requires-Dist: pandas==3.0.3; extra == 'dev'
+Requires-Dist: polars-runtime-32==1.42.1; extra == 'dev'
+Requires-Dist: polars==1.42.1; extra == 'dev'
+Requires-Dist: pyarrow==23.0.1; extra == 'dev'
+Requires-Dist: pytest-xdist>=3.6; extra == 'dev'
+Requires-Dist: pytest>=8.3; extra == 'dev'
+Requires-Dist: referencing<1,>=0.35; extra == 'dev'
+Requires-Dist: ruff>=0.9; extra == 'dev'
+Provides-Extra: hosted
+Requires-Dist: mostlyright-studio-client-v3==3.0.0; extra == 'hosted'
+Provides-Extra: local
+Requires-Dist: croniter==6.2.4; extra == 'local'
+Requires-Dist: cryptography<51,>=50.0.0; extra == 'local'
+Requires-Dist: et-xmlfile==2.0.0; extra == 'local'
+Requires-Dist: idna<4,>=3.10; extra == 'local'
+Requires-Dist: numpy==2.5.1; extra == 'local'
+Requires-Dist: openpyxl==3.1.5; extra == 'local'
+Requires-Dist: packaging<27,>=24; extra == 'local'
+Requires-Dist: pandas==3.0.3; extra == 'local'
+Requires-Dist: polars-runtime-32==1.42.1; extra == 'local'
+Requires-Dist: polars==1.42.1; extra == 'local'
+Requires-Dist: pyarrow==23.0.1; extra == 'local'
+Provides-Extra: local-search-neural
+Requires-Dist: anyio==4.14.2; extra == 'local-search-neural'
+Requires-Dist: certifi==2026.7.22; extra == 'local-search-neural'
+Requires-Dist: click==8.4.2; extra == 'local-search-neural'
+Requires-Dist: filelock==3.32.2; extra == 'local-search-neural'
+Requires-Dist: flatbuffers==25.12.19; extra == 'local-search-neural'
+Requires-Dist: fsspec==2026.7.0; extra == 'local-search-neural'
+Requires-Dist: h11==0.16.0; extra == 'local-search-neural'
+Requires-Dist: hf-xet==1.6.0; extra == 'local-search-neural'
+Requires-Dist: httpcore==1.0.9; extra == 'local-search-neural'
+Requires-Dist: httpx==0.28.1; extra == 'local-search-neural'
+Requires-Dist: huggingface-hub==1.27.0; extra == 'local-search-neural'
+Requires-Dist: idna==3.18; extra == 'local-search-neural'
+Requires-Dist: numpy==2.5.1; extra == 'local-search-neural'
+Requires-Dist: onnxruntime==1.28.0; extra == 'local-search-neural'
+Requires-Dist: packaging==26.3; extra == 'local-search-neural'
+Requires-Dist: protobuf==7.35.1; extra == 'local-search-neural'
+Requires-Dist: pyyaml==6.0.3; extra == 'local-search-neural'
+Requires-Dist: tokenizers==0.23.1; extra == 'local-search-neural'
+Requires-Dist: tqdm==4.70.0; extra == 'local-search-neural'
+Requires-Dist: typing-extensions==4.16.0; extra == 'local-search-neural'
+Provides-Extra: release
+Requires-Dist: croniter==6.2.4; extra == 'release'
+Requires-Dist: cryptography<51,>=50.0.0; extra == 'release'
+Requires-Dist: et-xmlfile==2.0.0; extra == 'release'
+Requires-Dist: idna<4,>=3.10; extra == 'release'
+Requires-Dist: jsonschema[format]<5,>=4.23; extra == 'release'
+Requires-Dist: numpy==2.5.1; extra == 'release'
+Requires-Dist: openpyxl==3.1.5; extra == 'release'
+Requires-Dist: packaging<27,>=24; extra == 'release'
+Requires-Dist: pandas==3.0.3; extra == 'release'
+Requires-Dist: polars-runtime-32==1.42.1; extra == 'release'
+Requires-Dist: polars==1.42.1; extra == 'release'
+Requires-Dist: pyarrow==23.0.1; extra == 'release'
+Provides-Extra: viewer
+Requires-Dist: nbconvert==7.17.1; extra == 'viewer'
+Requires-Dist: nbformat==5.10.4; extra == 'viewer'
+Description-Content-Type: text/markdown
+
+# mostlyright-data
+
+Mostly Right turns public weather, economic, financial, and satellite data into datasets for
+analysis and prediction. It keeps source links, timestamps, and check results with each dataset
+version.
+
+`mostlyright-data` is the command-line client. It starts hosted research and build work, follows
+run status, and downloads completed Parquet datasets.
+
+## Requirements
+
+Use CPython 3.13.
+
+## Install
+
+```bash
+python -m pip install mostlyright-data
+```
+
+The default install is the hosted client. It works on Linux, macOS, and Windows.
+
+## Sign in
+
+```bash
+mr-data login
+mr-data whoami
+```
+
+`login` stores a device key on the current machine. `whoami` checks that key.
+
+## Check the installation
+
+```bash
+mr-data --version
+mr-data --help
+```
+
+Use `mr-data research-open`, `mr-data research-status`, and `mr-data research-close` for hosted
+research sessions. `research-status` shows the current state.
+
+## Local tools
+
+The optional local tools require Linux:
+
+```bash
+python -m pip install "mostlyright-data[local]"
+```
+
+They build and check worker inputs. They are not required for the hosted client.
+
+## Documentation
+
+- [CLI guide](https://mostlyright.md/docs/guides/cli/)
+- [Research sessions](https://mostlyright.md/docs/guides/research/)
+- [Credentials](https://mostlyright.md/docs/guides/credentials/)
+- [Public datasets](https://mostlyright.md/docs/guides/use-public-datasets/)
+
+Use `mr-data --help` for the full command list and options. Commands that support
+`--json` write one JSON object.
+
+For local builds, `--input-root` and `--output` may use relative paths.
+
+## Limits and behavior
+
+- `mr-data review --local RUN_DIR` exits 1 when a check fails.
+- The weather Reader accepts one extracted GRIB2 record. It refuses a multi-record GFS or HRRR file.
+- Hosted credential-free acquisition is limited to one single-address public HTTPS object. Bounded collections require both local attestations.
+
+## Verification
+
+```bash
+make test-fast
+```
+
+`make test-fast` is non-authoritative and is useful while editing. `make verify` is required
+before review or merge.
+
+## Development
+
+```bash
+uv sync --locked --extra dev
+```
+
+The local test suite requires Linux for worker-boundary tests. The hosted client is tested on Linux,
+macOS, and Windows.

mostlyright_data-0.9.2/README.md

@@ -0,0 +1,85 @@
+# mostlyright-data
+
+Mostly Right turns public weather, economic, financial, and satellite data into datasets for
+analysis and prediction. It keeps source links, timestamps, and check results with each dataset
+version.
+
+`mostlyright-data` is the command-line client. It starts hosted research and build work, follows
+run status, and downloads completed Parquet datasets.
+
+## Requirements
+
+Use CPython 3.13.
+
+## Install
+
+```bash
+python -m pip install mostlyright-data
+```
+
+The default install is the hosted client. It works on Linux, macOS, and Windows.
+
+## Sign in
+
+```bash
+mr-data login
+mr-data whoami
+```
+
+`login` stores a device key on the current machine. `whoami` checks that key.
+
+## Check the installation
+
+```bash
+mr-data --version
+mr-data --help
+```
+
+Use `mr-data research-open`, `mr-data research-status`, and `mr-data research-close` for hosted
+research sessions. `research-status` shows the current state.
+
+## Local tools
+
+The optional local tools require Linux:
+
+```bash
+python -m pip install "mostlyright-data[local]"
+```
+
+They build and check worker inputs. They are not required for the hosted client.
+
+## Documentation
+
+- [CLI guide](https://mostlyright.md/docs/guides/cli/)
+- [Research sessions](https://mostlyright.md/docs/guides/research/)
+- [Credentials](https://mostlyright.md/docs/guides/credentials/)
+- [Public datasets](https://mostlyright.md/docs/guides/use-public-datasets/)
+
+Use `mr-data --help` for the full command list and options. Commands that support
+`--json` write one JSON object.
+
+For local builds, `--input-root` and `--output` may use relative paths.
+
+## Limits and behavior
+
+- `mr-data review --local RUN_DIR` exits 1 when a check fails.
+- The weather Reader accepts one extracted GRIB2 record. It refuses a multi-record GFS or HRRR file.
+- Hosted credential-free acquisition is limited to one single-address public HTTPS object. Bounded collections require both local attestations.
+
+## Verification
+
+```bash
+make test-fast
+```
+
+`make test-fast` is non-authoritative and is useful while editing. `make verify` is required
+before review or merge.
+
+## Development
+
+```bash
+uv sync --locked --extra dev
+```
+
+The local test suite requires Linux for worker-boundary tests. The hosted client is tested on Linux,
+macOS, and Windows.

{mostlyright_data-0.9.0 → mostlyright_data-0.9.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "mostlyright-data"
-version = "0.9.0"
+version = "0.9.2"
 description = "Mostly Right CLI for reproducible, reviewed Parquet datasets"
 readme = "README.md"
 requires-python = ">=3.13,<3.14"
@@ -14,6 +14,10 @@ requires-python = ">=3.13,<3.14"
 # runs the hosted commands there.
 dependencies = []
 
+[project.urls]
+Homepage = "https://mostlyright.md/"
+Documentation = "https://mostlyright.md/docs/guides/cli/"
+
 # Provenance-only resolution for the committed native Reader artifacts. Dependency groups do not
 # enter built wheel metadata or the base install closure; this keeps VENDOR.lock.json and uv.lock
 # independently binding the reviewed wheel bytes without making an unpublished wheel a runtime

{mostlyright_data-0.9.0 → mostlyright_data-0.9.2}/src/mostlyright/data_harness/hosted_session_worker.py

@@ -164,9 +164,17 @@ MAX_SESSION_DISPATCH_RESPONSE_BYTES = 64 * 1024
 #: stay no larger than the public request surface.  The child accepts exactly one such document.
 MAX_LEGACY_BOOTSTRAP_BYTES = MAX_START_REQUEST_BYTES
 #: Service start must not acknowledge until its fresh monitor has installed the cancellation
-#: handler that owns an inner capability child. This remains far below Studio's start request
-#: timeout while rejecting a stuck/interrupted bootstrap rather than accepting a zombie session.
-LEGACY_BOOTSTRAP_READY_TIMEOUT_SECONDS = 5.0
+#: handler that owns an inner capability child.  Fresh Python startup on a cold Cloud Run instance
+#: can legitimately take longer than the old five-second window, especially while the image and
+#: imports are being initialized.  Ten seconds remains below Studio's fifteen-second start request
+#: deadline, so Studio cannot settle a failed start before this Service acknowledges it.  It is
+#: still a fixed refusal bound: a stuck or interrupted bootstrap is torn down rather than accepted
+#: as a zombie session.
+# Production cold starts have reached just beyond ten seconds after the service
+# container is ready.  Keep the child bound below Studio's 15-second start-call
+# deadline, including the observed service-start overhead, while leaving room
+# for that measured import path.
+LEGACY_BOOTSTRAP_READY_TIMEOUT_SECONDS = 12.0
 #: Studio bounds the nonce at 32..128 URL-safe characters.
 MIN_NONCE_CHARS = 32
 MAX_NONCE_CHARS = 128
@@ -3013,6 +3021,11 @@ class _LegacySessionProcess:
                     "legacy bootstrap exceeds its byte limit",
                 )
             cls._write_all(write_descriptor, payload)
+            # The fresh bootstrap reads to EOF before it parses the payload. Close the writer
+            # before waiting for readiness; retaining it here makes the child wait for EOF while
+            # this parent waits for the child's acknowledgement.
+            os.close(write_descriptor)
+            write_descriptor = -1
             cls._await_ready(process)
         except (OSError, UnicodeError, SessionWorkerError, subprocess.SubprocessError) as error:
             if process is not None:
@@ -3055,7 +3068,16 @@ class _LegacySessionProcess:
             [stdout.fileno()], [], [], LEGACY_BOOTSTRAP_READY_TIMEOUT_SECONDS
         )
         if not readable:
-            raise OSError("legacy bootstrap monitor did not become ready in time")
+            # This path runs before the capability is leased.  Keep the diagnostic operationally
+            # useful while deliberately excluding the start invocation, nonce, and any capability.
+            LOGGER.warning(
+                "legacy bootstrap monitor missed its bounded readiness deadline (%.1fs)",
+                LEGACY_BOOTSTRAP_READY_TIMEOUT_SECONDS,
+            )
+            raise OSError(
+                "legacy bootstrap monitor did not become ready "
+                f"within {LEGACY_BOOTSTRAP_READY_TIMEOUT_SECONDS:.1f} seconds"
+            )
         raw = stdout.readline(MAX_SESSION_SUPERVISOR_MESSAGE_BYTES + 1)
         if len(raw) > MAX_SESSION_SUPERVISOR_MESSAGE_BYTES or not raw.endswith(b"\n"):
             raise OSError("legacy bootstrap monitor readiness is invalid")