ETLai 0.1.0__tar.gz

etlai-0.1.0/ETLai.egg-info/PKG-INFO

@@ -0,0 +1,193 @@
+Metadata-Version: 2.4
+Name: ETLai
+Version: 0.1.0
+Summary: Local, folder-driven Dagster application for CSV transformations with hot-folder sensors and Tkinter configuration UI.
+Author: Umangjeet Singh Pahwa
+License: MIT
+Project-URL: Repository, https://github.com/umang/ETLai
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Scientific/Engineering
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: dagster>=1.5
+Requires-Dist: dagster-webserver>=1.5
+Requires-Dist: pandas>=2.0
+Requires-Dist: faker>=18.0
+Requires-Dist: openpyxl>=3.1
+Dynamic: license-file
+
+# Air-Gapped Desktop Automation Engine
+
+A local Dagster application for folder-driven CSV transformations. Source files,
+saved column choices, generated outputs, and Dagster run storage remain on the
+machine.
+
+## How it works
+
+Each business pipeline has an inbox and a Dagster sensor. The sensor waits for
+the required number of stable files, moves them to staging, and launches a job.
+The job prepares atom parameters, executes the transformation, then moves source
+files to `processed/` or `rejected/`.
+
+An **atom** is a reusable transformation module with an
+`execute(params_json) -> result_json` interface. Atoms are domain-agnostic —
+they have no opinion on which column or business context. They do read and write
+files, so they are not side-effect-free.
+
+A **business pipeline** applies an atom to specific domain data with its own
+folder lifecycle, sensor, and saved configuration.
+
+## Registered jobs
+
+| Job | Files needed | Description |
+|-----|-------------|-------------|
+| `vlookup_rollnumber` | 2 CSV | Left-join on user-chosen columns (first run: Tkinter picker, then saved) |
+| `groupby_religion` | 1 CSV | Group by user-chosen column with counts, sorted descending |
+| `mock_generator` | 1+ CSV | Generate synthetic data from file headers using Faker |
+| `vlookup_then_groupby` | 2 CSV | Composite: VLOOKUP → GroupBy in a single job |
+
+Each job has a same-named sensor. Sensors must be enabled in Dagster UI before
+they process inbox files.
+
+## Installation
+
+Python 3.10+ required.
+
+```bash
+python3 -m venv .venv
+source .venv/bin/activate       # Windows: .venv\Scripts\activate
+python -m pip install -r requirements.txt
+```
+
+Tkinter must be available in the Python installation for first-run config.
+
+## Run
+
+```bash
+DAGSTER_HOME="$(pwd)" dagster dev -m definitions
+```
+
+Open `http://localhost:3000`, enable the desired sensor, and place files in
+`pipelines/<job-name>/inbox/`.
+
+Manual execution without UI:
+
+```bash
+DAGSTER_HOME="$(pwd)" dagster job execute -m definitions -j vlookup_rollnumber
+```
+
+## Source modules
+
+### Entry points
+
+| File | Role |
+|------|------|
+| `definitions.py` | Registers all jobs and sensors with Dagster |
+| `pipeline.py` | Builds single-atom business jobs via the factory |
+| `runners/composite.py` | Defines the `vlookup_then_groupby` composite job |
+
+### `logic/atoms/`
+
+| Module | Purpose |
+|--------|---------|
+| `vlookup.py` | Left join two CSVs on specified columns with dtype validation |
+| `groupby.py` | Group by one column, count, sort descending |
+| `mock_generate.py` | Infer Faker generators from headers, write synthetic CSVs |
+
+### `runners/`
+
+| Module | Purpose |
+|--------|---------|
+| `pipeline_factory.py` | Assembles load → preprocess → execute Dagster jobs |
+| `atom_runner.py` | Reusable load/execute ops with processed/rejected/notification lifecycle |
+| `ops.py` | First-run config ops for VLOOKUP, GroupBy, and mock generator |
+| `composite.py` | VLOOKUP → GroupBy chain with shared config |
+
+### `sensors/`
+
+| Module | Purpose |
+|--------|---------|
+| `hot_folder_sensor.py` | Per-pipeline sensor factory: stability check, staging, sweeper |
+
+### `helpers/`
+
+| Module | Purpose |
+|--------|---------|
+| `folders.py` | `PipelineFolders` class — per-pipeline directory lifecycle |
+| `config_store.py` | JSON config read/write per pipeline |
+| `column_picker.py` | Tkinter UI for VLOOKUP join/output column selection |
+| `notifier.py` | macOS/Windows toast notifications, opens output folder on success |
+| `file_picker.py` | Standalone Tkinter file dialog (not used by registered jobs) |
+| `mock_generator.py` | Standalone CLI mock generator (separate from the atom/job) |
+
+### Runtime data
+
+- `pipelines/<job>/` — inbox, staging, processed, rejected, output, config.json
+- `dagster.yaml` — SQLite storage configuration
+- `dagster_storage/` — Dagster run history (persists across restarts)
+
+## File and configuration lifecycle
+
+```text
+inbox/ → sensor stability check → staging/ → job runs
+                                    ├→ processed/ + output/       (success)
+                                    └→ rejected/ + *.error.txt    (failure)
+```
+
+Non-matching files (not `*.csv`/`*.xlsx`) are swept to `rejected/` after 3
+minutes with an error description.
+
+### Config init pattern
+
+1. First run: pre-process op detects no `config.json` → shows Tkinter UI → saves
+2. Future runs: loads config silently, fully automated
+3. To reconfigure: delete `pipelines/<job>/config.json`
+
+### Saved config fields
+
+| Job | Fields |
+|-----|--------|
+| `vlookup_rollnumber` | `left_column`, `right_column`, `left_output_columns`, `right_output_columns` |
+| `groupby_religion` | `group_column` |
+| `vlookup_then_groupby` | All VLOOKUP fields + `group_column` |
+| `mock_generator` | None |
+
+## Important behavior
+
+- Two-file jobs: lexical filename order determines left/right assignment.
+- Atom failures and config cancellation move staged files to `rejected/`.
+- Unexpected crashes may leave files in `staging/` — recover manually.
+- Notifications are best-effort. Linux has no implementation.
+- On success, macOS/Windows opens the output folder immediately (not a click
+  action).
+
+## Adding a new business pipeline
+
+1. Create or reuse an atom in `logic/atoms/` with `execute(params_json) -> str`
+2. Create a pre-process op in `runners/ops.py` (with config init pattern)
+3. Register in `pipeline.py`:
+   ```python
+   my_pipeline = build_business_pipeline(
+       pipeline_name="my_pipeline",
+       atom_module=my_atom,
+       atom_label="My Pipeline",
+       pre_process_op=my_pre_process_op,
+   )
+   ```
+4. Add sensor in `definitions.py`:
+   ```python
+   my_sensor = build_hot_folder_sensor("my_pipeline", "my_pipeline", min_files=1)
+   ```
+5. Folders are created automatically on import.
+
+## Development status
+
+No automated tests, lint, or pinned dependency versions. Validate by loading
+definitions and exercising jobs with disposable CSV inputs.

etlai-0.1.0/ETLai.egg-info/SOURCES.txt

@@ -0,0 +1,31 @@
+LICENSE
+README.md
+pyproject.toml
+ETLai.egg-info/PKG-INFO
+ETLai.egg-info/SOURCES.txt
+ETLai.egg-info/dependency_links.txt
+ETLai.egg-info/requires.txt
+ETLai.egg-info/top_level.txt
+helpers/README.md
+helpers/__init__.py
+helpers/column_picker.py
+helpers/config_store.py
+helpers/file_picker.py
+helpers/folders.py
+helpers/mock_generator.py
+helpers/notifier.py
+logic/README.md
+logic/__init__.py
+logic/atoms/README.md
+logic/atoms/__init__.py
+logic/atoms/groupby.py
+logic/atoms/vlookup.py
+runners/README.md
+runners/__init__.py
+runners/atom_runner.py
+runners/composite.py
+runners/ops.py
+runners/pipeline_factory.py
+sensors/README.md
+sensors/__init__.py
+sensors/hot_folder_sensor.py

etlai-0.1.0/ETLai.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

etlai-0.1.0/ETLai.egg-info/requires.txt

@@ -0,0 +1,5 @@
+dagster>=1.5
+dagster-webserver>=1.5
+pandas>=2.0
+faker>=18.0
+openpyxl>=3.1

etlai-0.1.0/ETLai.egg-info/top_level.txt

@@ -0,0 +1,4 @@
+helpers
+logic
+runners
+sensors

etlai-0.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Umangjeet Singh Pahwa
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

etlai-0.1.0/PKG-INFO

@@ -0,0 +1,193 @@
+Metadata-Version: 2.4
+Name: ETLai
+Version: 0.1.0
+Summary: Local, folder-driven Dagster application for CSV transformations with hot-folder sensors and Tkinter configuration UI.
+Author: Umangjeet Singh Pahwa
+License: MIT
+Project-URL: Repository, https://github.com/umang/ETLai
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Scientific/Engineering
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: dagster>=1.5
+Requires-Dist: dagster-webserver>=1.5
+Requires-Dist: pandas>=2.0
+Requires-Dist: faker>=18.0
+Requires-Dist: openpyxl>=3.1
+Dynamic: license-file
+
+# Air-Gapped Desktop Automation Engine
+
+A local Dagster application for folder-driven CSV transformations. Source files,
+saved column choices, generated outputs, and Dagster run storage remain on the
+machine.
+
+## How it works
+
+Each business pipeline has an inbox and a Dagster sensor. The sensor waits for
+the required number of stable files, moves them to staging, and launches a job.
+The job prepares atom parameters, executes the transformation, then moves source
+files to `processed/` or `rejected/`.
+
+An **atom** is a reusable transformation module with an
+`execute(params_json) -> result_json` interface. Atoms are domain-agnostic —
+they have no opinion on which column or business context. They do read and write
+files, so they are not side-effect-free.
+
+A **business pipeline** applies an atom to specific domain data with its own
+folder lifecycle, sensor, and saved configuration.
+
+## Registered jobs
+
+| Job | Files needed | Description |
+|-----|-------------|-------------|
+| `vlookup_rollnumber` | 2 CSV | Left-join on user-chosen columns (first run: Tkinter picker, then saved) |
+| `groupby_religion` | 1 CSV | Group by user-chosen column with counts, sorted descending |
+| `mock_generator` | 1+ CSV | Generate synthetic data from file headers using Faker |
+| `vlookup_then_groupby` | 2 CSV | Composite: VLOOKUP → GroupBy in a single job |
+
+Each job has a same-named sensor. Sensors must be enabled in Dagster UI before
+they process inbox files.
+
+## Installation
+
+Python 3.10+ required.
+
+```bash
+python3 -m venv .venv
+source .venv/bin/activate       # Windows: .venv\Scripts\activate
+python -m pip install -r requirements.txt
+```
+
+Tkinter must be available in the Python installation for first-run config.
+
+## Run
+
+```bash
+DAGSTER_HOME="$(pwd)" dagster dev -m definitions
+```
+
+Open `http://localhost:3000`, enable the desired sensor, and place files in
+`pipelines/<job-name>/inbox/`.
+
+Manual execution without UI:
+
+```bash
+DAGSTER_HOME="$(pwd)" dagster job execute -m definitions -j vlookup_rollnumber
+```
+
+## Source modules
+
+### Entry points
+
+| File | Role |
+|------|------|
+| `definitions.py` | Registers all jobs and sensors with Dagster |
+| `pipeline.py` | Builds single-atom business jobs via the factory |
+| `runners/composite.py` | Defines the `vlookup_then_groupby` composite job |
+
+### `logic/atoms/`
+
+| Module | Purpose |
+|--------|---------|
+| `vlookup.py` | Left join two CSVs on specified columns with dtype validation |
+| `groupby.py` | Group by one column, count, sort descending |
+| `mock_generate.py` | Infer Faker generators from headers, write synthetic CSVs |
+
+### `runners/`
+
+| Module | Purpose |
+|--------|---------|
+| `pipeline_factory.py` | Assembles load → preprocess → execute Dagster jobs |
+| `atom_runner.py` | Reusable load/execute ops with processed/rejected/notification lifecycle |
+| `ops.py` | First-run config ops for VLOOKUP, GroupBy, and mock generator |
+| `composite.py` | VLOOKUP → GroupBy chain with shared config |
+
+### `sensors/`
+
+| Module | Purpose |
+|--------|---------|
+| `hot_folder_sensor.py` | Per-pipeline sensor factory: stability check, staging, sweeper |
+
+### `helpers/`
+
+| Module | Purpose |
+|--------|---------|
+| `folders.py` | `PipelineFolders` class — per-pipeline directory lifecycle |
+| `config_store.py` | JSON config read/write per pipeline |
+| `column_picker.py` | Tkinter UI for VLOOKUP join/output column selection |
+| `notifier.py` | macOS/Windows toast notifications, opens output folder on success |
+| `file_picker.py` | Standalone Tkinter file dialog (not used by registered jobs) |
+| `mock_generator.py` | Standalone CLI mock generator (separate from the atom/job) |
+
+### Runtime data
+
+- `pipelines/<job>/` — inbox, staging, processed, rejected, output, config.json
+- `dagster.yaml` — SQLite storage configuration
+- `dagster_storage/` — Dagster run history (persists across restarts)
+
+## File and configuration lifecycle
+
+```text
+inbox/ → sensor stability check → staging/ → job runs
+                                    ├→ processed/ + output/       (success)
+                                    └→ rejected/ + *.error.txt    (failure)
+```
+
+Non-matching files (not `*.csv`/`*.xlsx`) are swept to `rejected/` after 3
+minutes with an error description.
+
+### Config init pattern
+
+1. First run: pre-process op detects no `config.json` → shows Tkinter UI → saves
+2. Future runs: loads config silently, fully automated
+3. To reconfigure: delete `pipelines/<job>/config.json`
+
+### Saved config fields
+
+| Job | Fields |
+|-----|--------|
+| `vlookup_rollnumber` | `left_column`, `right_column`, `left_output_columns`, `right_output_columns` |
+| `groupby_religion` | `group_column` |
+| `vlookup_then_groupby` | All VLOOKUP fields + `group_column` |
+| `mock_generator` | None |
+
+## Important behavior
+
+- Two-file jobs: lexical filename order determines left/right assignment.
+- Atom failures and config cancellation move staged files to `rejected/`.
+- Unexpected crashes may leave files in `staging/` — recover manually.
+- Notifications are best-effort. Linux has no implementation.
+- On success, macOS/Windows opens the output folder immediately (not a click
+  action).
+
+## Adding a new business pipeline
+
+1. Create or reuse an atom in `logic/atoms/` with `execute(params_json) -> str`
+2. Create a pre-process op in `runners/ops.py` (with config init pattern)
+3. Register in `pipeline.py`:
+   ```python
+   my_pipeline = build_business_pipeline(
+       pipeline_name="my_pipeline",
+       atom_module=my_atom,
+       atom_label="My Pipeline",
+       pre_process_op=my_pre_process_op,
+   )
+   ```
+4. Add sensor in `definitions.py`:
+   ```python
+   my_sensor = build_hot_folder_sensor("my_pipeline", "my_pipeline", min_files=1)
+   ```
+5. Folders are created automatically on import.
+
+## Development status
+
+No automated tests, lint, or pinned dependency versions. Validate by loading
+definitions and exercising jobs with disposable CSV inputs.

etlai-0.1.0/README.md

@@ -0,0 +1,168 @@
+# Air-Gapped Desktop Automation Engine
+
+A local Dagster application for folder-driven CSV transformations. Source files,
+saved column choices, generated outputs, and Dagster run storage remain on the
+machine.
+
+## How it works
+
+Each business pipeline has an inbox and a Dagster sensor. The sensor waits for
+the required number of stable files, moves them to staging, and launches a job.
+The job prepares atom parameters, executes the transformation, then moves source
+files to `processed/` or `rejected/`.
+
+An **atom** is a reusable transformation module with an
+`execute(params_json) -> result_json` interface. Atoms are domain-agnostic —
+they have no opinion on which column or business context. They do read and write
+files, so they are not side-effect-free.
+
+A **business pipeline** applies an atom to specific domain data with its own
+folder lifecycle, sensor, and saved configuration.
+
+## Registered jobs
+
+| Job | Files needed | Description |
+|-----|-------------|-------------|
+| `vlookup_rollnumber` | 2 CSV | Left-join on user-chosen columns (first run: Tkinter picker, then saved) |
+| `groupby_religion` | 1 CSV | Group by user-chosen column with counts, sorted descending |
+| `mock_generator` | 1+ CSV | Generate synthetic data from file headers using Faker |
+| `vlookup_then_groupby` | 2 CSV | Composite: VLOOKUP → GroupBy in a single job |
+
+Each job has a same-named sensor. Sensors must be enabled in Dagster UI before
+they process inbox files.
+
+## Installation
+
+Python 3.10+ required.
+
+```bash
+python3 -m venv .venv
+source .venv/bin/activate       # Windows: .venv\Scripts\activate
+python -m pip install -r requirements.txt
+```
+
+Tkinter must be available in the Python installation for first-run config.
+
+## Run
+
+```bash
+DAGSTER_HOME="$(pwd)" dagster dev -m definitions
+```
+
+Open `http://localhost:3000`, enable the desired sensor, and place files in
+`pipelines/<job-name>/inbox/`.
+
+Manual execution without UI:
+
+```bash
+DAGSTER_HOME="$(pwd)" dagster job execute -m definitions -j vlookup_rollnumber
+```
+
+## Source modules
+
+### Entry points
+
+| File | Role |
+|------|------|
+| `definitions.py` | Registers all jobs and sensors with Dagster |
+| `pipeline.py` | Builds single-atom business jobs via the factory |
+| `runners/composite.py` | Defines the `vlookup_then_groupby` composite job |
+
+### `logic/atoms/`
+
+| Module | Purpose |
+|--------|---------|
+| `vlookup.py` | Left join two CSVs on specified columns with dtype validation |
+| `groupby.py` | Group by one column, count, sort descending |
+| `mock_generate.py` | Infer Faker generators from headers, write synthetic CSVs |
+
+### `runners/`
+
+| Module | Purpose |
+|--------|---------|
+| `pipeline_factory.py` | Assembles load → preprocess → execute Dagster jobs |
+| `atom_runner.py` | Reusable load/execute ops with processed/rejected/notification lifecycle |
+| `ops.py` | First-run config ops for VLOOKUP, GroupBy, and mock generator |
+| `composite.py` | VLOOKUP → GroupBy chain with shared config |
+
+### `sensors/`
+
+| Module | Purpose |
+|--------|---------|
+| `hot_folder_sensor.py` | Per-pipeline sensor factory: stability check, staging, sweeper |
+
+### `helpers/`
+
+| Module | Purpose |
+|--------|---------|
+| `folders.py` | `PipelineFolders` class — per-pipeline directory lifecycle |
+| `config_store.py` | JSON config read/write per pipeline |
+| `column_picker.py` | Tkinter UI for VLOOKUP join/output column selection |
+| `notifier.py` | macOS/Windows toast notifications, opens output folder on success |
+| `file_picker.py` | Standalone Tkinter file dialog (not used by registered jobs) |
+| `mock_generator.py` | Standalone CLI mock generator (separate from the atom/job) |
+
+### Runtime data
+
+- `pipelines/<job>/` — inbox, staging, processed, rejected, output, config.json
+- `dagster.yaml` — SQLite storage configuration
+- `dagster_storage/` — Dagster run history (persists across restarts)
+
+## File and configuration lifecycle
+
+```text
+inbox/ → sensor stability check → staging/ → job runs
+                                    ├→ processed/ + output/       (success)
+                                    └→ rejected/ + *.error.txt    (failure)
+```
+
+Non-matching files (not `*.csv`/`*.xlsx`) are swept to `rejected/` after 3
+minutes with an error description.
+
+### Config init pattern
+
+1. First run: pre-process op detects no `config.json` → shows Tkinter UI → saves
+2. Future runs: loads config silently, fully automated
+3. To reconfigure: delete `pipelines/<job>/config.json`
+
+### Saved config fields
+
+| Job | Fields |
+|-----|--------|
+| `vlookup_rollnumber` | `left_column`, `right_column`, `left_output_columns`, `right_output_columns` |
+| `groupby_religion` | `group_column` |
+| `vlookup_then_groupby` | All VLOOKUP fields + `group_column` |
+| `mock_generator` | None |
+
+## Important behavior
+
+- Two-file jobs: lexical filename order determines left/right assignment.
+- Atom failures and config cancellation move staged files to `rejected/`.
+- Unexpected crashes may leave files in `staging/` — recover manually.
+- Notifications are best-effort. Linux has no implementation.
+- On success, macOS/Windows opens the output folder immediately (not a click
+  action).
+
+## Adding a new business pipeline
+
+1. Create or reuse an atom in `logic/atoms/` with `execute(params_json) -> str`
+2. Create a pre-process op in `runners/ops.py` (with config init pattern)
+3. Register in `pipeline.py`:
+   ```python
+   my_pipeline = build_business_pipeline(
+       pipeline_name="my_pipeline",
+       atom_module=my_atom,
+       atom_label="My Pipeline",
+       pre_process_op=my_pre_process_op,
+   )
+   ```
+4. Add sensor in `definitions.py`:
+   ```python
+   my_sensor = build_hot_folder_sensor("my_pipeline", "my_pipeline", min_files=1)
+   ```
+5. Folders are created automatically on import.
+
+## Development status
+
+No automated tests, lint, or pinned dependency versions. Validate by loading
+definitions and exercising jobs with disposable CSV inputs.

etlai-0.1.0/helpers/README.md

@@ -0,0 +1,43 @@
+# helpers/
+
+Cross-cutting utilities for file management, config, UI, and notifications.
+
+## `folders.py`
+
+`PipelineFolders(pipeline_name)` — per-pipeline directory lifecycle:
+- Properties: `.inbox`, `.staging`, `.processed`, `.rejected`, `.output`,
+  `.config_path`
+- Methods: `.ensure()`, `.move_to_staging()`, `.move_to_processed()`,
+  `.move_to_rejected(file_paths, reason)`, `.list_inbox_files(pattern)`,
+  `.output_path(filename)`
+
+## `config_store.py`
+
+Per-pipeline JSON config persistence:
+- `load_config(folders)` — Returns saved dict or None
+- `save_config(folders, config)` — Writes `config.json`
+- `config_exists(folders)` — Boolean check
+
+## `column_picker.py`
+
+Tkinter dialog for VLOOKUP join/output column selection. Shows columns from
+both files with dtypes, validates type compatibility on join columns, and
+allows multi-select for output columns from both sides.
+
+## `notifier.py`
+
+Best-effort native OS notifications:
+- macOS: AppleScript via `osascript`
+- Windows: PowerShell toast API
+- Opens output folder immediately on success (not a click action)
+- Fails silently on unsupported platforms (Linux)
+
+## `file_picker.py`
+
+Standalone Tkinter file-selection helper. Not used by registered Dagster jobs
+(the sensor provides files). Available for manual/ad-hoc use.
+
+## `mock_generator.py`
+
+Standalone two-file CLI mock-data generator. This is separate from the
+registered `mock_generator` atom/job — it predates the Dagster integration.