jerry-thomas 0.3.0__py3-none-any.whl → 1.0.1__py3-none-any.whl

jerry_thomas-0.3.0.dist-info/METADATA

@@ -1,502 +0,0 @@
-Metadata-Version: 2.4
-Name: jerry-thomas
-Version: 0.3.0
-Summary: Jerry-Thomas: a stream-first, plugin-friendly data pipeline (mixology-themed CLI)
-Author: Anders Skott Lind
-License: MIT
-Requires-Python: >=3.10
-Description-Content-Type: text/markdown
-License-File: LICENSE
-Requires-Dist: numpy<3.0,>=1.24
-Requires-Dist: pydantic>=2.0
-Requires-Dist: PyYAML>=5.4
-Requires-Dist: tqdm>=4.0
-Requires-Dist: jinja2>=3.0
-Provides-Extra: ml
-Requires-Dist: pandas>=2.0; extra == "ml"
-Requires-Dist: torch>=2.0; extra == "ml"
-Dynamic: license-file
-
-# Datapipeline Runtime
-
-Jerry Thomas is a time-series-first data pipeline runtime. It turns declarative
-YAML projects into iterators that stream records, engineered features, and
-model-ready vectors. The CLI lets you preview every stage, build deterministic
-artifacts, inspect coverage, and scaffold plugins for custom loaders, parsers,
-transforms, and filters.
-
-> **Core assumptions**
->
-> - Every record carries a timezone-aware `time` attribute and a numeric
->   `value`.
-> - Grouping is purely temporal. Dimensional splits belong in `partition_by`.
-
----
-
-## Why You Might Use It
-
-- Materialize canonical time-series datasets from disparate sources.
-- Preview and debug each stage of the pipeline without writing ad-hoc scripts.
-- Enforce coverage/quality gates and publish artifacts (expected IDs, scaler
-  stats) for downstream ML teams.
-- Extend the runtime with entry-point driven plugins for domain-specific I/O or
-  feature engineering.
-- Consume vectors directly from Python via iterators, Pandas DataFrames, or
-  `torch.utils.data.Dataset`.
-
----
-
-## Quick Start
-
-```bash
-# 1. Install in editable mode (with optional dev extras for testing).
-pip install -e .[dev]
-
-# 2. Bootstrap a project (scaffolds configs, plugin package, and templates).
-jerry plugin init --name my_datapipeline --out .
-
-# 3. Create a source & domain scaffold, then declare a canonical stream.
-jerry source add --provider demo --dataset weather --transport fs --format csv
-jerry domain add --domain weather
-# (edit config/contracts/<alias>.yaml to point at your mapper and policies)
-
-# 4. Configure dataset/postprocess/build files in config/datasets/<name>/.
-#    Then preview the pipeline and serve a few vectors:
-jerry serve --project config/datasets/default/project.yaml --stage 2 --limit 5
-jerry serve --project config/datasets/default/project.yaml --output print --limit 3
-
-# 5. Inspect coverage and build artifacts:
-jerry inspect report --project config/datasets/default/project.yaml
-jerry build --project config/datasets/default/project.yaml
-```
-
-The skeleton project in `src/datapipeline/templates/plugin_skeleton/` mirrors the
-paths expected by the CLI. Copy it or run `jerry plugin init` to get a ready-made
-layout with `config/`, `src/<package>/`, and entry-point stubs.
-
----
-
-## Pipeline Architecture
-
-```text
-raw source ──▶ canonical stream ──▶ record stage ──▶ feature stage ──▶ vector stage
-```
-
-1. **Raw sources** pair a loader with a parser. Loaders fetch bytes (file system,
-   HTTP, synthetic generators). Parsers turn those bytes into typed DTOs.
-   Register them via entry points (`loaders`, `parsers`) and declaratively wire
-   them in `config/sources/*.yaml`.
-2. **Canonical streams** decorate raw sources with mappers and per-stream
-   policies. Contract files under `config/contracts/` define record transforms,
-   feature transforms, sort hints, and partitioning.
-3. **Record stage** applies canonical policies to DTOs, turning them into
-   `TemporalRecord` instances (tz-aware timestamp + numeric value).
-4. **Feature stage** wraps records into `FeatureRecord`s, handles per-feature
-   sorting, optional scaling, and sequence windows (`FeatureRecordSequence`).
-5. **Vector stage** merges all feature streams, buckets them using `group_by`
-   cadence (e.g., `1h`), and emits `(group_key, Vector)` pairs ready for
-   downstream consumers.
-
-The runtime (`src/datapipeline/runtime.py`) hosts registries for sources,
-transforms, artifacts, and postprocess rules. The CLI constructs lightweight
-`PipelineContext` objects to build iterators without mutating global state.
-
----
-
-## Configuration Files
-
-All project configuration lives under `config/datasets/<name>/` by default.
-
-### `project.yaml`
-
-```yaml
-version: 1
-name: default
-paths:
-  streams: ../../contracts
-  sources: ../../sources
-  dataset: dataset.yaml
-  postprocess: postprocess.yaml
-  artifacts: ../../build/datasets/${project_name}
-  build: build.yaml
-  run: run.yaml
-globals:
-  start_time: 2021-01-01T00:00:00Z
-  end_time: 2023-01-03T23:00:00Z
-  split:
-    mode: hash # hash | time
-    key: group # group | feature:<id>
-    seed: 42
-    ratios: { train: 0.8, val: 0.1, test: 0.1 }
-```
-
-- `name` provides a stable identifier you can reuse inside config files via `${project_name}`.
-- `paths.*` are resolved relative to the project file unless absolute; they also support `${var}` interpolation.
-- `globals` provide values for `${var}` interpolation across YAML files. Datetime
-  values are normalized to strict UTC `YYYY-MM-DDTHH:MM:SSZ`.
-- `split` config defines how labels are assigned; the active label is selected by `run.yaml` or CLI `--keep`.
-- `paths.run` may point to a single file (default) or a directory. When it is a directory,
-  every `*.yaml` file inside is treated as a run config; `jerry serve` executes them
-  sequentially in alphabetical order unless you pass `--run <name>` (filename stem).
-- Label names are free-form: match whatever keys you declare in `split.ratios` (hash) or `split.labels` (time).
-
-### `run.yaml`
-
-```yaml
-version: 1
-keep: train # set to any label defined in globals.split (null disables filtering)
-output: print # override to 'stream' or a .pt path for binary dumps
-limit: 100 # cap vectors per serve run (null = unlimited)
-include_targets: false
-throttle_ms: null # sleep between vectors (milliseconds)
-log_level: INFO # DEBUG=progress bars, INFO=spinner, WARNING=quiet (null inherits CLI)
-```
-
-- `keep` selects the currently served split. This file is referenced by `project.paths.run`.
-- `output`, `limit`, `include_targets`, `throttle_ms`, and `log_level` provide defaults for `jerry serve`; CLI flags still win per invocation.
-- Override `keep` (and other fields) per invocation via `jerry serve ... --keep val` etc.
-- To manage multiple runs, point `project.paths.run` at a folder (e.g., `config/datasets/default/runs/`)
-  and drop additional `*.yaml` files there. `jerry serve` will run each file in order; pass
-  `--run train` to execute only `runs/train.yaml`.
-
-### `config/sources/<alias>.yaml`
-
-Each file defines a loader/parser pair exposed under `<alias>` (also the
-`source_id` the rest of the pipeline references).
-
-```yaml
-source_id: demo_weather
-loader:
-  entrypoint: demo.csv_loader
-  args:
-    path: data/weather.csv
-parser:
-  entrypoint: demo.weather_parser
-  args:
-    timezone: UTC
-```
-
-### `config/contracts/<alias>.yaml`
-
-Canonical stream contracts describe how the runtime should map and prepare a
-source. `alias` normally matches the source alias; use folders to organize by
-domain.
-
-```yaml
-source_id: demo_weather
-stream_id: demo_weather
-
-mapper:
-  entrypoint: weather.domain.mapper
-  args: {}
-
-partition_by: station
-sort_batch_size: 50000
-
-record:
-  - filter: { operator: ge, field: time, comparand: "${start_time}" }
-  - filter: { operator: lt, field: time, comparand: "${end_time}" }
-  - floor_time: { resolution: 10m }
-
-stream:
-  - ensure_ticks: { tick: 10m }
-  - granularity: { mode: mean }
-  - fill: { statistic: median, window: 6, min_samples: 2 }
-
-debug:
-  - lint: { mode: warn, tick: 10m }
-```
-
-- `record`: ordered record-level transforms (filters, floor/lag, custom
-  transforms registered under the `record` entry-point group).
-- `stream`: transforms applied after feature wrapping, still per base feature.
-- `debug`: instrumentation-only transforms (linters, assertions).
-- `partition_by`: optional keys used to suffix feature IDs (e.g., `temp__station=XYZ`).
-- `sort_batch_size`: chunk size used by the in-memory sorter when normalizing
-  order before stream transforms.
-
-### `dataset.yaml`
-
-Defines which canonical streams become features/targets and the vector bucketing.
-
-```yaml
-group_by: 1h
-
-features:
-  - id: temp_c
-    record_stream: demo_weather
-    scale: true
-    sequence: { size: 6, stride: 1, tick: 10m }
-
-targets:
-  - id: precip
-    record_stream: demo_weather
-```
-
-- `group_by` controls the cadence for vector partitioning (accepts `Xm|min|Xh`
-  — minutes or hours).
-- `scale: true` inserts the standard scaler feature transform (requires scaler
-  stats artifact or inline statistics).
-- `sequence` emits `FeatureRecordSequence` windows (size, stride, optional
-  cadence enforcement via `tick`).
-
-### `postprocess.yaml`
-
-Project-scoped vector transforms that run after assembly and before serving.
-
-```yaml
-- drop_missing:
-    required: [temp_c__station=001]
-    min_coverage: 0.95
-- fill_constant: { value: 0.0 }
-- fill_history:
-    statistic: median
-    window: 48
-    min_samples: 6
-- fill_horizontal:
-    statistic: mean
-    min_samples: 2
-```
-
-- Vector transforms rely on artifacts (expected IDs, scaler stats) to decide
-  which features should be present.
-- When no transforms are configured the stream passes through unchanged.
-
-### `build.yaml`
-
-Declares which artifacts the build step should materialize.
-
-```yaml
-version: 1
-partitioned_ids:
-  output: expected.txt
-  include_targets: false
-scaler:
-  enabled: true
-  output: scaler.pkl
-  include_targets: false
-  split_label: train
-```
-
-- `expected.txt` lists every fully partitioned feature ID observed in the latest
-  run (used by vector postprocess transforms).
-- `scaler.pkl` is a pickled standard scaler fitted on the requested split.
-
----
-
-## CLI Reference
-
-All commands live under the `jerry` entry point (`src/datapipeline/cli/app.py`).
-Pass `--help` on any command for flags.
-
-### Preview Stages
-
-- `jerry serve --project <project.yaml> --stage <0-7> --limit N [--log-level LEVEL]`
-  - Stage 0: raw DTOs
-  - Stage 1: domain `TemporalRecord`s
-  - Stage 2: record transforms applied
-  - Stage 3: feature records (before sort/regularization)
-  - Stage 4: feature regularization (post stream transforms)
-  - Stage 5: feature transforms/sequence outputs
-  - Stage 6: vectors assembled (no postprocess)
-  - Stage 7: vectors + postprocess transforms
-  - Use `--log-level DEBUG` for progress bars, `--log-level INFO` for spinner + prints, or the default (`WARNING`) for minimal output.
-- `jerry serve --project <project.yaml> --output print|stream|path.pt|path.csv|path.jsonl.gz --limit N [--include-targets] [--log-level LEVEL] [--run name]`
-  - Applies postprocess transforms and optional dataset split before emitting.
-  - Set `--log-level DEBUG` (or set `run.yaml` -> `log_level: DEBUG`) to reuse the tqdm progress bars when previewing stages.
-  - When `project.paths.run` is a directory, add `--run val` (filename stem) to target a single config; otherwise every run file is executed sequentially.
-  - Argument precedence: CLI flags > run.yaml > built‑in defaults.
-
-### Build & Quality
-
-- `jerry inspect report --project <project.yaml> [--threshold 0.95] [--include-targets]`
-  - Prints coverage summary (keep/below lists) and writes `coverage.json` under
-    the artifacts directory.
-  - Add `--matrix csv|html` to persist an availability matrix.
-- `jerry inspect partitions --project <project.yaml> [--include-targets]`
-  - Writes discovered partition suffixes to `partitions.json`.
-- `jerry inspect expected --project <project.yaml> [--include-targets]`
-  - Writes the full set of observed feature IDs to `expected.txt`.
-- `jerry build --project <project.yaml> [--force]`
-  - Regenerates artifacts declared in `build.yaml` if configuration hash changed.
-
-### Scaffolding & Reference
-
-- `jerry plugin init --name <package> --out <dir>`
-  - Generates a plugin project (pyproject, package skeleton, config templates).
-- `jerry source add --provider <name> --dataset <slug> --transport fs|url|synthetic --format csv|json|json-lines`
-  - Creates loader/parser stubs, updates entry points, and drops a matching
-    source YAML.
-- `jerry domain add --domain <name>`
-  - Adds a `domains/<name>/` package with a `model.py` stub.
-- `jerry filter create --name <identifier>`
-  - Scaffolds an entry-point-ready filter (helpful for custom record predicates).
-- `jerry list sources|domains`
-  - Introspect configured source aliases or domain packages.
-
----
-
-## Transform & Filter Library
-
-### Record Filters (`config/contracts[].record`)
-
-- Binary comparisons: `eq`, `ne`, `lt`, `le`, `gt`, `ge` (timezone-aware for ISO
-  or datetime literals).
-- Membership: `in`, `nin`.
-  ```yaml
-  - filter: { operator: ge, field: time, comparand: "${start_time}" }
-  - filter: { operator: in, field: station, comparand: [a, b, c] }
-  ```
-
-### Record Transforms
-
-- `floor_time`: snap timestamps down to the nearest resolution (`10m`, `1h`, …).
-- `lag`: add lagged copies of records (see `src/datapipeline/transforms/record/lag.py` for options).
-
-### Stream (Feature) Transforms
-
-- `ensure_ticks`: backfill missing ticks with `value=None` records to enforce a
-  strict cadence.
-- `granularity`: merge duplicate timestamps using `first|last|mean|median`.
-- `fill`: rolling statistic-based imputation within each feature stream.
-- Custom transforms can be registered under the `stream` entry-point group.
-
-### Feature Transforms
-
-- `scale`: wraps `StandardScalerTransform`. Read statistics from the build
-  artifact or accept inline `statistics`.
-  ```yaml
-  scale:
-    with_mean: true
-    with_std: true
-    statistics:
-      temp_c__station=001: { mean: 10.3, std: 2.1 }
-  ```
-
-### Sequence Transforms
-
-- `sequence`: sliding window generator (`size`, `stride`, optional `tick` to
-  enforce gaps). Emits `FeatureRecordSequence` payloads with `.records`.
-
-### Vector (Postprocess) Transforms
-
-- `drop_missing`: drop vectors that do not meet required IDs or coverage ratio.
-- `fill_constant`: seed absent IDs with a constant.
-- `fill_history`: impute using rolling statistics from prior vectors.
-- `fill_horizontal`: aggregate sibling partitions in the same timestamp.
-
-All transforms share a consistent entry-point signature and accept their config
-dict as keyword arguments. Register new ones in `pyproject.toml` under the
-appropriate group (`record`, `stream`, `feature`, `sequence`, `vector`,
-`filters`, `debug`).
-
----
-
-## Artifacts & Postprocess
-
-- `expected.txt`: newline-delimited full feature IDs. Required by drop/fill
-  transforms to know the target feature universe.
-- `scaler.pkl`: pickled standard scaler fitted on the configured split. Loaded
-  lazily by feature transforms at runtime.
-- Build state is tracked in `artifacts/build/state.json`; config hashes avoid
-  redundant runs.
-
-If a postprocess transform needs an artifact and it is missing, the runtime will
-raise a descriptive error suggesting `jerry build`.
-
----
-
-## Splitting & Serving
-
-If `project.globals.split` is present, `jerry serve` filters vectors at the
-end of the pipeline:
-
-- `mode: hash` – deterministic entity hash using either the group key or a
-  specified feature ID.
-- `mode: time` – boundary-based slicing using timestamp labels.
-- `run.keep` (or CLI `--keep`) selects the active slice; use any label name defined in your split config.
-
-The split configuration never mutates stored artifacts; it is only applied when
-serving vectors (either via CLI or the Python integrations).
-
----
-
-## Python Integrations
-
-`datapipeline.integrations.ml` demonstrates how to reuse the runtime from
-application code:
-
-- `VectorAdapter.from_project(project_yaml)` – bootstrap once, then stream
-  vectors or row dicts.
-- `stream_vectors(project_yaml, limit=...)` – iterator matching `jerry serve`.
-- `iter_vector_rows` / `collect_vector_rows` – handy for Pandas or custom sinks.
-- `dataframe_from_vectors` – eager helper that returns a Pandas DataFrame
-  (requires `pandas`).
-- `torch_dataset` – builds a `torch.utils.data.Dataset` that yields tensors. See
-  `examples/minimal_project/run_torch.py` for usage.
-
----
-
-## Extending the Runtime
-
-### Entry Points
-
-Register custom components in your plugin’s `pyproject.toml`:
-
-```toml
-[project.entry-points."datapipeline.loaders"]
-demo.csv_loader = "my_datapipeline.loaders.csv:CsvLoader"
-
-[project.entry-points."datapipeline.parsers"]
-demo.weather_parser = "my_datapipeline.parsers.weather:WeatherParser"
-
-[project.entry-points."datapipeline.mappers"]
-weather.domain.mapper = "my_datapipeline.mappers.weather:DomainMapper"
-
-[project.entry-points."datapipeline.stream"]
-weather.fill = "my_datapipeline.transforms.weather:CustomFill"
-```
-
-Loader, parser, mapper, and transform classes should provide a callable
-interface (usually `__call__`) matching the runtime expectations. Refer to the
-built-in implementations in `src/datapipeline/sources/`, `src/datapipeline/transforms/`,
-and `src/datapipeline/filters/`.
-
-### Scaffolding Helpers
-
-- `datapipeline.services.scaffold.plugin.scaffold_plugin` – invoked by
-  `jerry plugin init`.
-- `datapipeline.services.scaffold.source.create_source` – writes loader/parser
-  stubs and updates entry points.
-- `datapipeline.services.scaffold.domain.create_domain` – domain DTO skeleton.
-- `datapipeline.services.scaffold.filter.create_filter` – custom filter stub.
-- `datapipeline.services.scaffold.mappers.attach_source_to_domain` – helper for
-  programmatically wiring sources to domain mappers and emitting stream
-  contracts (useful in custom automation or tests).
-
----
-
-## Development Workflow
-
-- Install dependencies: `pip install -e .[dev]`.
-- Run tests: `pytest`.
-- When iterating on configs, use `jerry serve --stage <n>` to peek into problematic
-  stages.
-- After tuning transforms, refresh artifacts: `jerry build`.
-- Use `jerry inspect report --include-targets` to ensure targets meet coverage
-  gates before handing vectors to downstream consumers.
-
----
-
-## Additional Resources
-
-- `src/datapipeline/analysis/vector_analyzer.py` – quality metrics collected by
-  the inspect commands.
-- `src/datapipeline/pipeline/` – pure functions that wire each stage.
-- `src/datapipeline/services/bootstrap/` – runtime initialization and
-  registry population (see `core.py`).
-- `examples/minimal_project/` – runnable demo showing config layout and Torch
-  integration.
-
-Happy shipping! Build, inspect, and serve consistent time-series features with
-confidence.

jerry_thomas-0.3.0.dist-info/RECORD

@@ -1,139 +0,0 @@
-datapipeline/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-datapipeline/plugins.py,sha256=Y0QfI313t5_w_m1ayQVEuac3lJ4YR_OSIYZol35ZOTk,838
-datapipeline/runtime.py,sha256=zNZmJNA6OTDl9NXGJikm5pRM7EDGAFtv4B_ErzDxVsY,2484
-datapipeline/analysis/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-datapipeline/analysis/vector_analyzer.py,sha256=D6eDW0H55QGnWiULXJEirMjw6MeBcwjJ4zfS7M6tx98,175
-datapipeline/analysis/vector/collector.py,sha256=S-CHdKfWeRP7TuxQdi22z9tJ9rt8VtbN48uJa-dgvBg,10233
-datapipeline/analysis/vector/matrix.py,sha256=mbMJoj7rzf6wQMV-1HbS4OvriWcfkF40ox8bYet-0zg,17861
-datapipeline/analysis/vector/report.py,sha256=lPL1he-tf5MeO2TFPteOiFLov6c6JgEStsvwWFvjrao,11205
-datapipeline/build/__init__.py,sha256=XbuHhJzIRTNnOOJMYa_BHiA0P8yPrEk4tuVXnD2NQbI,109
-datapipeline/build/state.py,sha256=XsL2CtQl7x80OpE1SJE42D-ig3lBMPr7_HSlpq5xwU4,1826
-datapipeline/build/tasks.py,sha256=j2DzVTOjfkcEHb4HQA1fMC5_7o-4i_ffyv_fZMw71TE,6367
-datapipeline/cli/app.py,sha256=V2vuFfkOJemKEImN21uUENBYqTODNY_JcxS5WwTyZFE,16416
-datapipeline/cli/commands/build.py,sha256=N1ekzK4lO6W4Dfuw58jvOHtHsdMM8rjZqD5CgGVDSmo,1641
-datapipeline/cli/commands/domain.py,sha256=KZ0dn08k7wdjBWmJPgeQ0ogI7fQfhCmkcckIpTlMYD0,308
-datapipeline/cli/commands/filter.py,sha256=vhoCIETJNUJmiI37ZdBNaeJAm6O4AU_tveJxVj47S8A,307
-datapipeline/cli/commands/inspect.py,sha256=YXQmLAij3tkOlXp0TLioB3v_rBOpuH0ZeJPVLtI07Bk,7916
-datapipeline/cli/commands/link.py,sha256=vvGlVrEXQtqIcxETwMXvQq1NWVHuGsNCDD8VxIoV57c,4907
-datapipeline/cli/commands/list_.py,sha256=cCrx35CPNMyAaOVvVxkZje4DAx7K7HdGpngzZwhNn40,943
-datapipeline/cli/commands/plugin.py,sha256=9AKFnebNbUkuZ5RzFBAPfKp9NERwd4uhrhPctiPLtYM,331
-datapipeline/cli/commands/run.py,sha256=525eqyeT8of4jG6O41VHtfu-kJrpL1OYMUXMMqBElGE,9049
-datapipeline/cli/commands/source.py,sha256=U21LAOiX10dXSmBGDGC2gjJC_GoIQ8Pmbq_E6Gil-kE,822
-datapipeline/cli/commands/writers.py,sha256=oLlMutWH6xHYQjR_zP59dJw4tDGFgexgN8RubeAICYk,4287
-datapipeline/cli/visuals/__init__.py,sha256=3VAWQt8tEKN_mIyhPAKqBqBy38fulA_DFVtfIPWaCnM,297
-datapipeline/cli/visuals/labels.py,sha256=3Dk7RoCO4hyxpEiBV9ltmZd9Fr9i8lroxnBjNGbyO-M,2604
-datapipeline/cli/visuals/sources.py,sha256=Z6pjPq46n0KbDXJfmVpLxV-z7e9wQhJZ03nQREA8yKQ,4160
-datapipeline/config/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-datapipeline/config/build.py,sha256=HcwOjuIfwM9PbMPaHncfpIhhI1WvfXWvgeRvrJKt_5c,1999
-datapipeline/config/catalog.py,sha256=ezGSy2yFybnYZ2EHU9IjDdeFVHgHhQbePAHGXFPY9b8,1012
-datapipeline/config/postprocess.py,sha256=E9Ic_yPCXVOmCRpmsFsHVI1t8JdIrT9nNqHn9H7kPg0,415
-datapipeline/config/project.py,sha256=GymhVVRbL-k0l7POKCRFfRUt6OMIJGpxp0b2i5JbuNw,988
-datapipeline/config/run.py,sha256=6SLmXC_wFJtyNKKUa2L3ZclQ6DR_OdG0h9vq-ylfnsU,3726
-datapipeline/config/split.py,sha256=VFYRF6Fz5xLTqqxIt3RVGB4kwlnHH8CxjOddEAJYG5Q,1048
-datapipeline/config/dataset/dataset.py,sha256=Q9cb5QoDtyPb4pbD9mSTZcJmXQhdEWwDLS52xKAcqXg,562
-datapipeline/config/dataset/feature.py,sha256=2Hxz0FXZskLI4ICXhmlG6b1Vvxzh0Ql9e6BwjMRtzSs,346
-datapipeline/config/dataset/loader.py,sha256=C_2zewx6Hk1MsQQSwctZaH2d3KPDzjnq5L08g5JMdLU,650
-datapipeline/config/dataset/normalize.py,sha256=66yvPvbQyef6qQtyJOGTkAo5Q5ldDqpk-XIHx9hYH4c,825
-datapipeline/domain/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-datapipeline/domain/feature.py,sha256=7BOI4H458BKU8B9vqdfez7WOO1YKiF6lt0oy7PMbqrQ,295
-datapipeline/domain/record.py,sha256=Q-QjuR3FbQ01_vJ5LT9k8E40C-oYmOvWdbqtwtfIHPM,709
-datapipeline/domain/vector.py,sha256=apK1iu7tca2k2xgNGJAAQfKhirno7ZKZ3pDheKf9euM,1041
-datapipeline/filters/filters.py,sha256=dM6U-QpGCQQ4-CMBTJgWZp2zH2TVTk6uYOqGPC5NBCY,2649
-datapipeline/integrations/__init__.py,sha256=tjTLsIa6NRWKI05wjwPAUuXozDA-gP98SccFJ9lYHs8,410
-datapipeline/integrations/ml/__init__.py,sha256=oflJXnjQEn1Zv0Vho10mc2y3D6UkKusNZwE5yUtatb8,463
-datapipeline/integrations/ml/adapter.py,sha256=Xy4VXJ0pi5Ydsv_0MAYXpXGf_GY84SPwSseMfHsrVk8,3979
-datapipeline/integrations/ml/pandas_support.py,sha256=pSpSOFpstxD9gOwOCd-XuQ8bfkVPv6VRkxnfXTgJUWI,1305
-datapipeline/integrations/ml/rows.py,sha256=KyGl36LI7F-LhjZQnal0bhlR2TPB9gDchKtOwDyPgt8,2219
-datapipeline/integrations/ml/torch_support.py,sha256=RBuvEYD-YV-daC63yth6QIC4SnEWY_3KCV5ifvwHCmw,2799
-datapipeline/mappers/noop.py,sha256=L8bH1QVbLH-ogIam0ppYdx7KuWQ7Dj44lvD8tvNlY0Q,111
-datapipeline/mappers/synthetic/time.py,sha256=lt1pC0May6Y4E8bZO4sERm3D04_r-qv63Y5fwrtCaBQ,639
-datapipeline/parsers/identity.py,sha256=pdGuz0SSQGfySPpvZSnLgfTXTkC36x-7dQMMei3XhsU,321
-datapipeline/pipeline/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-datapipeline/pipeline/context.py,sha256=0dlm8hRZn6fqXvgGg0zydHaibKwUQoxBM6seh2lmfTU,1914
-datapipeline/pipeline/pipelines.py,sha256=7X0BpJQcxHG6UCE4mCVjAvD2adyqQyO4wp_K5GFfAbo,2882
-datapipeline/pipeline/split.py,sha256=tvyerkvjZBF4WKI8VKEmx8deE-SFIG1bSUMgj62gBWk,6122
-datapipeline/pipeline/stages.py,sha256=Ehx2TeuIWdMW1fGI3uG6cowMUhOFhw7j2RYdWaI49d0,5704
-datapipeline/pipeline/utils/keygen.py,sha256=wg7dBlHUp9az4BNwiiYZ84A1VIbohUKoqL7ShC73V60,1400
-datapipeline/pipeline/utils/memory_sort.py,sha256=iPWcD81xtZZ8SXEX_Ph-hUCnpKlJMFybbxPZVOQdSs8,735
-datapipeline/pipeline/utils/ordering.py,sha256=ZX18I7GFtvyMFJB96vWQpTOGwljjeJ6ppCg5a3Av3es,1450
-datapipeline/pipeline/utils/transform_utils.py,sha256=CfQGSzM4BCbX5NR1ERP-uRkSqG9dIQLvHnD1-vquCv8,3202
-datapipeline/registries/registry.py,sha256=MWWOHz2wT1oHQmovodtEreEuQhvH-i11Y2yXUUgZJhQ,641
-datapipeline/services/artifacts.py,sha256=qHDVeDdeMUlYPpe_g0xuo9xUl5Y_aqqFgIPK9ywP_yc,2847
-datapipeline/services/constants.py,sha256=LpV5KKCLWSxsKk0sR4236s1aF01oVCi1Xqn60tmtu0M,488
-datapipeline/services/entrypoints.py,sha256=NKcSbhGRtBLQXGf-TdujwbVSRH1zb5J-S2jxFPnk6HQ,2504
-datapipeline/services/factories.py,sha256=yod9bGp5ErfVbeDym4TC13lez84F8P3wnnyTRuJKBVs,1009
-datapipeline/services/paths.py,sha256=xHxos62Y2gjhLggrnrmRqPiLMseK10OX17NJjnVk8wE,966
-datapipeline/services/project_paths.py,sha256=_Td5sFLRqtfXBVwYxCeuBuh_3s5V8ymo3GPgLv4aXI8,3249
-datapipeline/services/bootstrap/__init__.py,sha256=Mc2w2S69kU1hnzCvsGMhFqyNoNMXPwQtxprAkGN-sYE,245
-datapipeline/services/bootstrap/config.py,sha256=rIhp0J8j4ZQIaJ_8Y7ioa8sYGj4O-e9Q6q8MG8L9RmU,4207
-datapipeline/services/bootstrap/core.py,sha256=M9c8DujiJlaWXYyIGdLda3YktKrM4riGCWEQcBW-1ag,6497
-datapipeline/services/scaffold/__init__.py,sha256=PaQNtYki9Kc7mQPnUtKDPU-rKohLHoXLvFVwdHdbXNM,68
-datapipeline/services/scaffold/domain.py,sha256=POYlFTRak3eDjB6u4DtjijGfNzpX3SrVw_BZYcH6G4E,918
-datapipeline/services/scaffold/filter.py,sha256=fVb6PokoDnJxDsA9nlaHp-rnZlOA4aAnVJ76jJQQBhs,1060
-datapipeline/services/scaffold/mappers.py,sha256=Kef04q2YA1hJ6OG_Fbc_SGVU9CrUwcRjpHEjFz1nt40,1993
-datapipeline/services/scaffold/plugin.py,sha256=KZM_mKZsKy_KyyiJjioBzzgD05tt1kCpP2QBF9JFQVU,1701
-datapipeline/services/scaffold/source.py,sha256=_kJO4Kpvlpr_93TzLa4JJd2B7nEErwE-xZa1RtVOPRc,5538
-datapipeline/services/scaffold/templates.py,sha256=B3YnZpFUZLynijJosTNxZQLXnPP_Y_t1RHqfI1lGOxU,634
-datapipeline/sources/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-datapipeline/sources/composed_loader.py,sha256=dUmdAYncL6IJqwLhKpQ3AVyFNyut-_K55isSTr-iDXo,1350
-datapipeline/sources/decoders.py,sha256=OEozYeV9bAA0yiWebPRvq_qKt0KXArk5jBJk_2sr7iI,1970
-datapipeline/sources/factory.py,sha256=J177Y1TJsAJZimd54bKkJl1c7-rq8e2cs04ie3tHY8E,2097
-datapipeline/sources/transports.py,sha256=yBfRLlZxxP4INQ5uFSR8zJecjUx_XlwU6KMQqJUYx44,2080
-datapipeline/sources/models/__init__.py,sha256=rS3Rc8L2eXSd2CW9ejRConk9pmV8Zv7QCz3ekSIdcLQ,397
-datapipeline/sources/models/base.py,sha256=MAUawd11fII-mxxuSPM4f6H1t1tbyZX_QWhoAgeYUcU,238
-datapipeline/sources/models/generator.py,sha256=9hzZ_Tvdc8w_VnFMny_C0lLvhjpPgW22y2aXJGKSvdw,661
-datapipeline/sources/models/loader.py,sha256=WWQhyatf0dr6pIgh2Dbtif1r09kbCKq9G4-EX4PN9Fg,1025
-datapipeline/sources/models/parser.py,sha256=Ts31aksHLDCw5ovF2D99w9g_j-NnEiZ8x0JHtUxmmXs,226
-datapipeline/sources/models/source.py,sha256=PBtbJVdyuRABPGFSwkyDaSmT1BuHk2faL-lUvRFpOAo,796
-datapipeline/sources/models/synthetic.py,sha256=FLF2Jvdc06VCriTCliThuQTUXd6NrXIQpksIL8gBIH8,288
-datapipeline/sources/synthetic/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-datapipeline/sources/synthetic/time/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-datapipeline/sources/synthetic/time/loader.py,sha256=KSkLbaN4snFsA8U-ZjPoiUlTHX7u6sTH6L5kN828sEQ,1540
-datapipeline/sources/synthetic/time/parser.py,sha256=d3GZMQ7L1Qi4LeEm7U3y0_pk0RdhskioQukYyqyoqic,343
-datapipeline/templates/plugin_skeleton/README.md,sha256=5bdNVR9gqtXf2gZlTaXFLJrWadsakqqS-L9JJuesBaY,5982
-datapipeline/templates/plugin_skeleton/pyproject.toml,sha256=0lmO5Aia9tB81Ez4SxP56DGisekx-palMmGCUzmAl4E,259
-datapipeline/templates/plugin_skeleton/config/contracts/time_hour_sin.synthetic.yaml,sha256=FrSy5PPbnEGmZcduRuAdj5UiA7YOpJKSGzgVrsVB5x0,575
-datapipeline/templates/plugin_skeleton/config/contracts/time_linear.synthetic.yaml,sha256=3qSV6HZOPRZjQnLZqx2VbQsNzpwhjoYxhRQIcBUDcp4,1250
-datapipeline/templates/plugin_skeleton/config/datasets/default/build.yaml,sha256=zl4WCMc5M4Mke0B-pJD2CsdOsdUUQkfuDFDW9JH6bYs,431
-datapipeline/templates/plugin_skeleton/config/datasets/default/dataset.yaml,sha256=3LOOVBY36ECCqUf88RyVctlaZyggoQZaJp7Pok6FsBg,297
-datapipeline/templates/plugin_skeleton/config/datasets/default/postprocess.yaml,sha256=q-9VhMjB2j-Le4vZ3pZEH-N98DZEzDqXI3RvJXW-ung,310
-datapipeline/templates/plugin_skeleton/config/datasets/default/project.yaml,sha256=jsujHRmJ1mxXFGtjxN_wWzHEMfxp8UGY5fvLitXflu4,686
-datapipeline/templates/plugin_skeleton/config/datasets/default/runs/run_test.yaml,sha256=9_ZU0Zd1DYZoE3kd_PMZ1xIs6cv4NvbgxA2LVTbRzaw,580
-datapipeline/templates/plugin_skeleton/config/datasets/default/runs/run_train.yaml,sha256=fXy9RnBG2I_8VeQrfgQ5aJibqtwi6H0o754xQGgLJ4s,582
-datapipeline/templates/plugin_skeleton/config/datasets/default/runs/run_val.yaml,sha256=nqZ_h5jrUhnCzgvhKyv8Wb3hSxfcGF7p5N1VLE15sg0,578
-datapipeline/templates/plugin_skeleton/config/sources/time_ticks.yaml,sha256=7tie6CqEmOOK8M629f2WZDWKaUPr9bePMC6Oj8RsqB8,190
-datapipeline/templates/plugin_skeleton/src/{{PACKAGE_NAME}}/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-datapipeline/templates/stubs/dto.py.j2,sha256=0tEtjGmCu1ddZee_fjmKyRrx357u3Yo6MKxXIpU8224,877
-datapipeline/templates/stubs/filter.py.j2,sha256=3LgRgAL_HRaENOOqQx8NdeM1AUy-T0rtHVTA7N2oWOs,466
-datapipeline/templates/stubs/loader_synthetic.py.j2,sha256=9SQBeTBGlZmKs6nSYBKd8nbOPcFHgDx17Mh8xOEQnvs,1285
-datapipeline/templates/stubs/mapper.py.j2,sha256=5pKaGUEiUWAgLU4fVTXdx4LNFsIQqI4LX5FYKcayxIY,762
-datapipeline/templates/stubs/parser.py.j2,sha256=KDWO_kDogHosqFt4MGki9QncEpdeD-KCkAWqrhLe7Lo,672
-datapipeline/templates/stubs/parser_custom.py.j2,sha256=0Nytq43JdTZoyRj-4Mz6HWdMTmOP3VlFuYOB_A_13Vg,580
-datapipeline/templates/stubs/record.py.j2,sha256=bzI8Jt0hcInLpn-IlyL2a8-q4VQY4gyZ2Z5BAMB70k4,660
-datapipeline/templates/stubs/source.yaml.j2,sha256=mPOfYD3Hyvaw-lSgJslfsP2VqOp7qsg_ePOLvyTeSRw,416
-datapipeline/transforms/filter.py,sha256=Jt8wTEIqWqe34s7GVVekcR8OdRozs317sj7Uw08GNOA,1433
-datapipeline/transforms/sequence.py,sha256=5i-0w1jQcSHy12rhztBhzyhJ2FdnVbD35NWcPXPi_kQ,3059
-datapipeline/transforms/utils.py,sha256=ts6dULY2Pc5fFs7AMd3goN4hDzQkv-6CDLdRH41lG9I,721
-datapipeline/transforms/vector.py,sha256=sXDNMlVcqpRFXcMw_oC5nEM1mMQa59AnZG6e_1PrnBQ,7609
-datapipeline/transforms/vector_utils.py,sha256=PcStTwRaaunONKZJuwv79bjdfaDcamLcwNLRHjZ5yXw,927
-datapipeline/transforms/debug/identity.py,sha256=6bwnEYhMBYw0YPrMccrZPXDOQM4r_-odsKo8Hhpbz10,2515
-datapipeline/transforms/debug/lint.py,sha256=6EBzGOfYjJbHzcZIIzVixlvW5RVr7liw6DieuWwxNUM,4057
-datapipeline/transforms/feature/model.py,sha256=gB-GP80_P7bzEKJFSM4leRke75yiD4-S5eJ1p8g3JU8,382
-datapipeline/transforms/feature/scaler.py,sha256=pcIb-dstNsZQ-T4ZsONkKwkg1fVWpJDV19FyQPM9HpM,5265
-datapipeline/transforms/record/floor_time.py,sha256=dKxLjnmBNJmDClPQfuKBEM_lrW-356v8XfQtLog5K2k,627
-datapipeline/transforms/record/lag.py,sha256=5wrPyVNFvidvdQddnK6ZeUOI5I8rfXEbzIg6tzKiJu4,536
-datapipeline/transforms/stream/ensure_ticks.py,sha256=Q0AwKuRY2nRIOUKoaAINeAWUEuoOzSh30Ug5k8296Kw,1170
-datapipeline/transforms/stream/fill.py,sha256=N_ybLUCvaMVvKsFP8-HcGuKqV9hXAnYmV7zyUB-Ugys,3500
-datapipeline/transforms/stream/granularity.py,sha256=PzHDGDwyn8P07BCbcFZaorS_7lbAbEdMLqD9Wy61y0M,3376
-datapipeline/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-datapipeline/utils/load.py,sha256=7BJIx11DPiEtDaLf-ZOClCQxJ7c8oFcc3JdcnGjl4ig,1327
-datapipeline/utils/paths.py,sha256=5Y5rhNbjTiybUHfq9VfRMJ4gUfN9UltonM-4MABEG8w,798
-datapipeline/utils/pickle_model.py,sha256=Uyd4AajInyTUpWfSJDDEGLinXeQkHjQUNnyla0owtA4,854
-datapipeline/utils/placeholders.py,sha256=epZQ7NifUWI7_7hZKGEkCBDOaMnN9LiqJdI2gvBAEgE,890
-datapipeline/utils/time.py,sha256=vOqa2arqwEqbDo-JWEhOFPMnI1E4Ib3i1L-Rt-cGH8c,1072
-jerry_thomas-0.3.0.dist-info/licenses/LICENSE,sha256=pkBMylAJF5yChHAkdxwFhEptLGx13i-XFEKh-Sh6DkM,1073
-jerry_thomas-0.3.0.dist-info/METADATA,sha256=_zSaetgFZLkGnfG3Gd8UDDWQBZAwKlup-VAN_I-aNJ0,18413
-jerry_thomas-0.3.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-jerry_thomas-0.3.0.dist-info/entry_points.txt,sha256=7GKSNCgwbzIqI3_LdbOro2eiAbBvbpoOxAuh2XcqBN0,1669
-jerry_thomas-0.3.0.dist-info/top_level.txt,sha256=N8aoNPdPyHefODO4YAm7tqTaUcw0e8LDcqycFTf8TbM,13
-jerry_thomas-0.3.0.dist-info/RECORD,,