domherre 0.1.1__tar.gz

domherre-0.1.1/.gitignore

@@ -0,0 +1,226 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[codz]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#   Usually these files are written by a python script from a template
+#   before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py.cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+# Pipfile.lock
+
+# UV
+#   Similar to Pipfile.lock, it is generally recommended to include uv.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+# uv.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+# poetry.lock
+# poetry.toml
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#   pdm recommends including project-wide configuration in pdm.toml, but excluding .pdm-python.
+#   https://pdm-project.org/en/latest/usage/project/#working-with-version-control
+# pdm.lock
+# pdm.toml
+.pdm-python
+.pdm-build/
+
+# pixi
+#   Similar to Pipfile.lock, it is generally recommended to include pixi.lock in version control.
+# pixi.lock
+#   Pixi creates a virtual environment in the .pixi directory, just like venv module creates one
+#   in the .venv directory. It is recommended not to include this directory in version control.
+.pixi
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# Redis
+*.rdb
+*.aof
+*.pid
+
+# RabbitMQ
+mnesia/
+rabbitmq/
+rabbitmq-data/
+
+# ActiveMQ
+activemq-data/
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.envrc
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#   JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#   be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#   and can be added to the global gitignore or merged into this file.  For a more nuclear
+#   option (not recommended) you can uncomment the following to ignore the entire idea folder.
+# .idea/
+
+# Abstra
+#   Abstra is an AI-powered process automation framework.
+#   Ignore directories containing user credentials, local state, and settings.
+#   Learn more at https://abstra.io/docs
+.abstra/
+
+# Visual Studio Code
+#   Visual Studio Code specific template is maintained in a separate VisualStudioCode.gitignore 
+#   that can be found at https://github.com/github/gitignore/blob/main/Global/VisualStudioCode.gitignore
+#   and can be added to the global gitignore or merged into this file. However, if you prefer, 
+#   you could uncomment the following to ignore the entire vscode folder
+# .vscode/
+# Temporary file for partial code execution
+tempCodeRunnerFile.py
+
+# Ruff stuff:
+.ruff_cache/
+
+# PyPI configuration file
+.pypirc
+
+# Marimo
+marimo/_static/
+marimo/_lsp/
+__marimo__/
+
+# Streamlit
+.streamlit/secrets.toml
+
+# macOS metadata
+.DS_Store
+
+# Model comparison output
+compareModels/results_*.txt
+compareModels/stats_file.txt

domherre-0.1.1/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Alexander Koutakis
+
+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.

domherre-0.1.1/PKG-INFO

@@ -0,0 +1,219 @@
+Metadata-Version: 2.5
+Name: domherre
+Version: 0.1.1
+Summary: Redact Swedish names and personnummer from polars DataFrames
+License: MIT License
+        
+        Copyright (c) 2026 Alexander Koutakis
+        
+        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.
+License-File: LICENSE
+Requires-Python: >=3.10
+Requires-Dist: polars>=1.0.0
+Provides-Extra: all
+Requires-Dist: gliner>=0.2.0; extra == 'all'
+Requires-Dist: pandas>=2.0.0; extra == 'all'
+Requires-Dist: presidio-analyzer>=2.2.0; extra == 'all'
+Requires-Dist: torch>=2.0.0; extra == 'all'
+Requires-Dist: transformers>=4.30.0; extra == 'all'
+Provides-Extra: gliner
+Requires-Dist: gliner>=0.2.0; extra == 'gliner'
+Provides-Extra: pandas
+Requires-Dist: pandas>=2.0.0; extra == 'pandas'
+Provides-Extra: presidio
+Requires-Dist: presidio-analyzer>=2.2.0; extra == 'presidio'
+Provides-Extra: transformers
+Requires-Dist: torch>=2.0.0; extra == 'transformers'
+Requires-Dist: transformers>=4.30.0; extra == 'transformers'
+Description-Content-Type: text/markdown
+
+# domherre
+
+Redacts Swedish personal names, personnummer and e-mail adresses from string columns in a polars DataFrame.
+
+- **Personnummer** — regex (date-shaped) + Luhn checksum.
+- **Names** — `KB/bert-base-swedish-cased-ner`, `PRS` entities only. 
+
+## Install
+
+```bash
+pip install domherre
+```
+
+## What you get
+
+```python
+import polars as pl
+import domherre as kt
+
+df = pl.DataFrame({
+    "id": [1, 2],
+    "note": [
+        "Engelbert Karlsson, personnummer 850101-0006, ringde idag.",
+        "Ingen känslig info här.",
+    ],
+})
+
+kt.redact(df, "note")
+# note -> note_redacted
+# "[NAMN], personnummer [PERSONNUMMER], ringde idag."
+```
+
+- Redacted columns are **renamed** with a suffix (default `_redacted`) so a redacted column
+- can never be mistaken for the source. The original column is replaced, not kept.
+- Unlisted columns are untouched. Raises if a target name already exists.
+
+Several columns at once:
+
+```python
+kt.redact(df, ["note", "kommentar"])
+```
+
+Single string:
+
+```python
+kt.redact_text("Anna Andersson, 850101-0006")
+```
+
+
+A one-time `PandasPerformanceWarning` fires per process. Silence it with
+`DOMHERRE_QUIET=1` if you have already made your peace with pandas.
+
+## Backend models
+
+Name detection is pluggable. Personnummer detection is always regex + Luhn and is
+not affected by the backend choice. This project defaults to kb-bert model as it has shown to be the formidable model for swedish name detection.
+
+| Backend | Install | Notes |
+|---|---|---|
+| `kb-bert` (default) | included | `KB/bert-base-swedish-cased-ner`. Swedish-only, trained on formal text. |
+| `gliner` | `domherre[gliner]` | Zero-shot, multilingual, CPU-optimised. Change `labels` to detect anything. |
+| `presidio` | `domherre[presidio]` | Full PII framework. Bring your own configured `AnalyzerEngine`. |
+| `none` | included | Skips name detection. Regex only. |
+
+```python
+kt.redact(df, "note", kt.Config(backend="gliner"))
+```
+
+### gliner
+
+```python
+config = kt.Config(
+    backend="gliner",
+    model="urchade/gliner_multi_pii-v1",   # default
+    labels=("person", "full name"),
+    min_score=0.5,
+    backend_options={"map_location": "cpu", "quantize": True},
+)
+```
+
+
+### Custom backends
+
+Anything with `load()` and `person_spans(texts) -> list[list[tuple[int, int]]]` works.
+A plain callable works too:
+
+```python
+def detector(texts: list[str]) -> list[list[tuple[int, int]]]:
+    return [[(m.start(), m.end()) for m in pattern.finditer(t)] for t in texts]
+
+kt.redact(df, "note", kt.Config(backend=detector))
+```
+
+Register by name to make it selectable like a builtin:
+
+```python
+kt.register_backend("mine", MyBackend)
+kt.available_backends()   # ['gliner', 'kb-bert', 'mine', 'none', 'presidio']
+```
+
+Backends are cached per (name, config), so a Deployment loads each model once.
+
+## Config
+
+```python
+config = kt.Config(
+    name_tag="[NAMN]",
+    personnummer_tag="[PERSONNUMMER]",
+    suffix="_redacted",
+    backend="kb-bert",       # or "gliner", "presidio", "none", or your own
+    model=None,              # backend default, or a local path: "/models/ner"
+    local_only=False,        # True = never hit the network, fail loudly if not cached
+    num_threads=2,           # pin to your k8s cpu limit
+    batch_size=32,
+    min_score=0.0,           # raise to cut false-positive names, at the cost of misses
+    redact_names=True,
+    redact_personnummer=True,
+    validate_personnummer=True,   # False = redact on shape alone, catches typo'd numbers
+    enumerate_names=False,        # [NAMN_1], [NAMN_2] instead of flat [NAMN]
+    extra_patterns={},            # {"[EPOST]": r"\S+@\S+\.\w+"}
+)
+
+kt.redact(df, "note", config)
+```
+
+### enumerate_names
+
+Distinguishes people within a single string, so downstream analysis can still tell
+that two mentions refer to the same person:
+
+```
+"[NAMN_1] mailade [NAMN_2] om [NAMN_1]s ärende"
+```
+
+Numbering is per string, not per DataFrame. `[NAMN_1]` in row 1 and row 2 are not
+necessarily the same person — that would need identity resolution, which this does not do.
+
+## Service deployment
+
+Domherre is suited as deployemnts that can be run as services, we thus support models pre-baked into images.
+Load the model at startup so the first request doesn't pay for it:
+
+```python
+kt.load(backend="kb-bert", num_threads=2, local_only=True)
+```
+
+Bake the weights into the image and set `local_only=True`. Otherwise every cold pod
+pulls ~440MB from HuggingFace, and a network blip becomes a runtime failure
+in the middle of a batch instead of a clear error at boot.
+
+```dockerfile
+RUN python -c "from transformers import pipeline; \
+    pipeline('token-classification', model='KB/bert-base-swedish-cased-ner')"
+```
+
+```yaml
+resources:
+  requests: { cpu: "1", memory: "2Gi" }
+  limits:   { cpu: "2", memory: "3Gi" }
+```
+
+Set `num_threads` to match the CPU limit. Torch otherwise reads the node's core count,
+not the cgroup limit, and thrashes threads on a large node.
+
+## Personnummer formats detected
+
+`YYMMDD-XXXX`, `YYMMDD+XXXX`, `YYMMDDXXXX`, `YYYYMMDD-XXXX`, `YYYYMMDDXXXX`,
+plus samordningsnummer (day + 60). All must pass the Luhn check unless
+`validate_personnummer=False`.
+
+Not detected: reservnummer/interimsnummer (region-specific, no fixed national format),
+and personnummer with a typo'd control digit — set `validate_personnummer=False`
+to catch those, at the cost of redacting unrelated 10 and 12-digit numbers.
+

domherre-0.1.1/README.md

@@ -0,0 +1,174 @@
+# domherre
+
+Redacts Swedish personal names, personnummer and e-mail adresses from string columns in a polars DataFrame.
+
+- **Personnummer** — regex (date-shaped) + Luhn checksum.
+- **Names** — `KB/bert-base-swedish-cased-ner`, `PRS` entities only. 
+
+## Install
+
+```bash
+pip install domherre
+```
+
+## What you get
+
+```python
+import polars as pl
+import domherre as kt
+
+df = pl.DataFrame({
+    "id": [1, 2],
+    "note": [
+        "Engelbert Karlsson, personnummer 850101-0006, ringde idag.",
+        "Ingen känslig info här.",
+    ],
+})
+
+kt.redact(df, "note")
+# note -> note_redacted
+# "[NAMN], personnummer [PERSONNUMMER], ringde idag."
+```
+
+- Redacted columns are **renamed** with a suffix (default `_redacted`) so a redacted column
+- can never be mistaken for the source. The original column is replaced, not kept.
+- Unlisted columns are untouched. Raises if a target name already exists.
+
+Several columns at once:
+
+```python
+kt.redact(df, ["note", "kommentar"])
+```
+
+Single string:
+
+```python
+kt.redact_text("Anna Andersson, 850101-0006")
+```
+
+
+A one-time `PandasPerformanceWarning` fires per process. Silence it with
+`DOMHERRE_QUIET=1` if you have already made your peace with pandas.
+
+## Backend models
+
+Name detection is pluggable. Personnummer detection is always regex + Luhn and is
+not affected by the backend choice. This project defaults to kb-bert model as it has shown to be the formidable model for swedish name detection.
+
+| Backend | Install | Notes |
+|---|---|---|
+| `kb-bert` (default) | included | `KB/bert-base-swedish-cased-ner`. Swedish-only, trained on formal text. |
+| `gliner` | `domherre[gliner]` | Zero-shot, multilingual, CPU-optimised. Change `labels` to detect anything. |
+| `presidio` | `domherre[presidio]` | Full PII framework. Bring your own configured `AnalyzerEngine`. |
+| `none` | included | Skips name detection. Regex only. |
+
+```python
+kt.redact(df, "note", kt.Config(backend="gliner"))
+```
+
+### gliner
+
+```python
+config = kt.Config(
+    backend="gliner",
+    model="urchade/gliner_multi_pii-v1",   # default
+    labels=("person", "full name"),
+    min_score=0.5,
+    backend_options={"map_location": "cpu", "quantize": True},
+)
+```
+
+
+### Custom backends
+
+Anything with `load()` and `person_spans(texts) -> list[list[tuple[int, int]]]` works.
+A plain callable works too:
+
+```python
+def detector(texts: list[str]) -> list[list[tuple[int, int]]]:
+    return [[(m.start(), m.end()) for m in pattern.finditer(t)] for t in texts]
+
+kt.redact(df, "note", kt.Config(backend=detector))
+```
+
+Register by name to make it selectable like a builtin:
+
+```python
+kt.register_backend("mine", MyBackend)
+kt.available_backends()   # ['gliner', 'kb-bert', 'mine', 'none', 'presidio']
+```
+
+Backends are cached per (name, config), so a Deployment loads each model once.
+
+## Config
+
+```python
+config = kt.Config(
+    name_tag="[NAMN]",
+    personnummer_tag="[PERSONNUMMER]",
+    suffix="_redacted",
+    backend="kb-bert",       # or "gliner", "presidio", "none", or your own
+    model=None,              # backend default, or a local path: "/models/ner"
+    local_only=False,        # True = never hit the network, fail loudly if not cached
+    num_threads=2,           # pin to your k8s cpu limit
+    batch_size=32,
+    min_score=0.0,           # raise to cut false-positive names, at the cost of misses
+    redact_names=True,
+    redact_personnummer=True,
+    validate_personnummer=True,   # False = redact on shape alone, catches typo'd numbers
+    enumerate_names=False,        # [NAMN_1], [NAMN_2] instead of flat [NAMN]
+    extra_patterns={},            # {"[EPOST]": r"\S+@\S+\.\w+"}
+)
+
+kt.redact(df, "note", config)
+```
+
+### enumerate_names
+
+Distinguishes people within a single string, so downstream analysis can still tell
+that two mentions refer to the same person:
+
+```
+"[NAMN_1] mailade [NAMN_2] om [NAMN_1]s ärende"
+```
+
+Numbering is per string, not per DataFrame. `[NAMN_1]` in row 1 and row 2 are not
+necessarily the same person — that would need identity resolution, which this does not do.
+
+## Service deployment
+
+Domherre is suited as deployemnts that can be run as services, we thus support models pre-baked into images.
+Load the model at startup so the first request doesn't pay for it:
+
+```python
+kt.load(backend="kb-bert", num_threads=2, local_only=True)
+```
+
+Bake the weights into the image and set `local_only=True`. Otherwise every cold pod
+pulls ~440MB from HuggingFace, and a network blip becomes a runtime failure
+in the middle of a batch instead of a clear error at boot.
+
+```dockerfile
+RUN python -c "from transformers import pipeline; \
+    pipeline('token-classification', model='KB/bert-base-swedish-cased-ner')"
+```
+
+```yaml
+resources:
+  requests: { cpu: "1", memory: "2Gi" }
+  limits:   { cpu: "2", memory: "3Gi" }
+```
+
+Set `num_threads` to match the CPU limit. Torch otherwise reads the node's core count,
+not the cgroup limit, and thrashes threads on a large node.
+
+## Personnummer formats detected
+
+`YYMMDD-XXXX`, `YYMMDD+XXXX`, `YYMMDDXXXX`, `YYYYMMDD-XXXX`, `YYYYMMDDXXXX`,
+plus samordningsnummer (day + 60). All must pass the Luhn check unless
+`validate_personnummer=False`.
+
+Not detected: reservnummer/interimsnummer (region-specific, no fixed national format),
+and personnummer with a typo'd control digit — set `validate_personnummer=False`
+to catch those, at the cost of redacting unrelated 10 and 12-digit numbers.
+

domherre-0.1.1/pyproject.toml

@@ -0,0 +1,39 @@
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[project]
+name = "domherre"
+version = "0.1.1"
+readme = "README.md"
+license = { file = "LICENSE" }
+description = "Redact Swedish names and personnummer from polars DataFrames"
+requires-python = ">=3.10"
+dependencies = [
+    "polars>=1.0.0",
+]
+
+[project.optional-dependencies]
+transformers = ["transformers>=4.30.0", "torch>=2.0.0"]
+gliner = ["gliner>=0.2.0"]
+presidio = ["presidio-analyzer>=2.2.0"]
+pandas = ["pandas>=2.0.0"]
+all = ["domherre[transformers,gliner,presidio,pandas]"]
+
+[dependency-groups]
+dev = [
+    "pytest>=7.0.0",
+    "pandas>=2.0.0",
+    "mypy>=1.10.0",
+    "ruff>=0.6.0",
+]
+
+[tool.hatch.build.targets.wheel]
+packages = ["src/domherre"]
+
+[tool.hatch.build.targets.sdist]
+include = ["src/", "tests/", "README.md", "LICENSE"]
+
+[tool.pytest.ini_options]
+testpaths = ["tests"]
+pythonpath = ["src"]