catchfly 0.0.1__tar.gz

catchfly-0.0.1/.gitignore

@@ -0,0 +1,207 @@
+# 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
+
+# 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/
+
+# Ruff stuff:
+.ruff_cache/
+
+# PyPI configuration file
+.pypirc
+
+# Cursor
+#  Cursor is an AI-powered code editor. `.cursorignore` specifies files/directories to
+#  exclude from AI features like autocomplete and code analysis. Recommended for sensitive data
+#  refer to https://docs.cursor.com/context/ignore-files
+.cursorignore
+.cursorindexingignore
+
+# Marimo
+marimo/_static/
+marimo/_lsp/
+__marimo__/

catchfly-0.0.1/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 amichalski2
+
+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.

catchfly-0.0.1/PKG-INFO

@@ -0,0 +1,195 @@
+Metadata-Version: 2.4
+Name: catchfly
+Version: 0.0.1
+Summary: From chaos to categories. Schema discovery + term normalization for unstructured text data.
+Project-URL: Homepage, https://github.com/silene-systems/catchfly
+Project-URL: Documentation, https://github.com/silene-systems/catchfly
+Project-URL: Repository, https://github.com/silene-systems/catchfly
+Project-URL: Issues, https://github.com/silene-systems/catchfly/issues
+Author-email: Adrian Michalski <adrian@silene.systems>
+License-Expression: Apache-2.0
+License-File: LICENSE
+Keywords: clustering,deduplication,llm,nlp,normalization,rare-diseases,schema-discovery,text-classification
+Classifier: Development Status :: 1 - Planning
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Science/Research
+Classifier: License :: OSI Approved :: Apache Software 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: Programming Language :: Python :: 3.13
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Topic :: Text Processing :: Linguistic
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+
+<p align="center">
+  <h1 align="center">catchfly</h1>
+  <p align="center"><strong>From chaos to categories.</strong></p>
+  <p align="center">Schema discovery + term normalization for unstructured text data.</p>
+  <p align="center">
+    <a href="https://pypi.org/project/catchfly/"><img src="https://img.shields.io/pypi/v/catchfly?color=6B2D8B" alt="PyPI"></a>
+    <a href="https://github.com/silene-systems/catchfly/blob/main/LICENSE"><img src="https://img.shields.io/badge/license-Apache--2.0-green" alt="License"></a>
+    <a href="https://www.python.org/"><img src="https://img.shields.io/pypi/pyversions/catchfly" alt="Python"></a>
+  </p>
+</p>
+
+---
+
+> *Named after [Silene](https://en.wikipedia.org/wiki/Silene) (catchfly) — plants that secrete a sticky substance to capture insects. Catchfly captures scattered terms and groups them into canonical categories.*
+>
+> *Part of the [Silene Systems](https://silene.systems) ecosystem for rare disease research.*
+
+## The problem
+
+You extracted 3,000 mentions from 200 scientific papers. Or an LLM classified 3 million emails into 4,000 labels. Now you have:
+
+- **No schema** — you don't know what categories should exist
+- **Duplicates everywhere** — "miglustat" and "Zavesca" are the same drug
+- **Ambiguous boundaries** — is "cognitive decline" the same as "cognitive impairment"?
+- **Related but distinct entities** — "ALT" and "AST" are similar but clinically different
+
+No existing tool does both **schema discovery** and **term normalization** as a composable Python library.
+
+## What catchfly does
+
+**Two operations, one pipeline:**
+
+**Discover** — *"What categories exist in my data?"*
+
+```python
+from catchfly import Resolver
+
+resolver = Resolver(embed_provider="gemini", llm_provider="openai/gpt-5.4")
+
+schema = resolver.discover(
+    mentions=["miglustat", "splenomegaly", "ALT: 120", "ataxia", ...],
+    contexts={"miglustat": ["Patient received miglustat 200mg daily"], ...},
+)
+
+# schema.categories → ["Treatments", "Symptoms", "Lab Values", ...]
+# schema.examples  → {"Treatments": ["miglustat", "arimoclomol"], ...}
+
+schema.rename("Lab Values", "Laboratory Findings")
+schema.merge("Symptoms", "Clinical Signs")
+```
+
+**Normalize** — *"Which terms belong where, and which are synonyms?"*
+
+```python
+result = resolver.normalize(
+    mentions=all_mentions,
+    schema=schema,
+    contexts=contexts,
+)
+
+# result.groups    → [NormGroup(canonical="miglustat", members=["Zavesca", "NB-DNJ"])]
+# result.ambiguous → [AmbiguousPair("cognitive decline", "cognitive impairment")]
+```
+
+**Or all-in-one:**
+
+```python
+result = resolver.resolve(mentions, contexts=contexts)
+```
+
+## Self-improving prompts
+
+Give catchfly 20–50 labeled examples and it optimizes its own prompts for your domain. Inspired by [GEPA](https://github.com/gepa-ai/gepa), implemented from scratch with zero external dependencies.
+
+```python
+optimized = resolver.optimize(
+    mentions=mentions,
+    contexts=contexts,
+    ground_truth={"miglustat": "miglustat", "Zavesca": "miglustat", "ALT": "ALT", "AST": "AST"},
+    iterations=20,  # ~$1–2, ~15 min
+)
+
+result = optimized.resolve(new_mentions)  # 20–30% better accuracy on your domain
+```
+
+## How it works
+
+Four components, composable as a pipeline:
+
+| Component | What it does | Cost |
+|---|---|---|
+| **EmbeddingSimilarity** | Pre-filter: finds candidate pairs via cosine similarity | Embedding API only |
+| **LLMGrouper** | Core engine: LLM analyzes clusters, proposes categories/synonyms with 4-way relation typing (synonym / hierarchy / related / distinct) | LLM API |
+| **LLMClassifier** | Propagation: assigns remaining mentions to approved categories via few-shot classification | LLM API |
+| **UserSeeded** | User guidance: seed mappings override and anchor discovery + normalization | Embedding only |
+
+## Tiered quality
+
+| Tier | What runs | Cost / 1K mentions | Accuracy | Best for |
+|---|---|---|---|---|
+| **Tier 1** (free) | Embedding only | $0 (local models) | 60–70% | Exploration |
+| **Tier 2** (standard) | Embedding + LLM | $0.15–0.50 | 82–90% | Production |
+| **Tier 3** (optimized) | Tier 2 + optimize() | $1–3 one-time | 90–95%+ | Systematic reviews |
+
+## Use cases
+
+- **Medical literature** — normalize symptoms, drugs, genetic variants from systematic reviews
+- **Email categorization** — collapse 4,000 LLM-generated labels into 150 clean groups
+- **E-commerce tags** — build taxonomy from 50,000 user-generated product tags
+- **Any domain** — if you have messy text labels, catchfly cleans them up
+
+## Provider support
+
+```python
+# Embeddings
+Resolver(embed_provider="gemini")       # Google Gemini Embedding 2
+Resolver(embed_provider="openai")        # OpenAI text-embedding-3
+Resolver(embed_provider="local")         # sentence-transformers (offline, free)
+Resolver(embed_provider=my_function)     # any callable
+
+# LLM
+Resolver(llm_provider="openai/gpt-5.4")
+Resolver(llm_provider="gemini/flash")
+Resolver(llm_provider=my_function)       # any callable
+```
+
+## Evaluation
+
+```python
+ground_truth = {"miglustat": "miglustat", "Zavesca": "miglustat", "ALT": "ALT", "AST": "AST"}
+metrics = resolver.evaluate(result, ground_truth)
+# → precision, recall, false_merges, missed_merges, category_accuracy
+```
+
+## Installation
+
+```bash
+pip install catchfly
+```
+
+> **v0.0.1 is a name reservation.** Active development underway. First functional release (v0.1.0) expected Q2 2026.
+
+## Part of the Silene ecosystem
+
+| Project | What it is |
+|---|---|
+| **[Silene Systems](https://silene.systems)** | Computational phenotyping platform for rare diseases |
+| **Campion** | Agentic batch literature extraction platform (SaaS) |
+| **catchfly** | Schema discovery + term normalization library (open source, this repo) |
+
+> *[Silene tomentosa](https://en.wikipedia.org/wiki/Silene_tomentosa) (Gibraltar campion) is one of the rarest plants in the world — thought extinct in 1992, rediscovered in 1994 on the Rock of Gibraltar. Like rare diseases, it hides in plain sight, waiting for someone to look carefully enough.*
+
+## License
+
+Apache-2.0
+
+## Citation
+
+If you use catchfly in academic work:
+
+```bibtex
+@software{catchfly2026,
+  author = {Michalski, Adrian},
+  title = {catchfly: Schema discovery and category normalization for unstructured text data},
+  year = {2026},
+  url = {https://github.com/silene-systems/catchfly},
+  license = {Apache-2.0}
+}
+``

catchfly-0.0.1/README.md

@@ -0,0 +1,169 @@
+<p align="center">
+  <h1 align="center">catchfly</h1>
+  <p align="center"><strong>From chaos to categories.</strong></p>
+  <p align="center">Schema discovery + term normalization for unstructured text data.</p>
+  <p align="center">
+    <a href="https://pypi.org/project/catchfly/"><img src="https://img.shields.io/pypi/v/catchfly?color=6B2D8B" alt="PyPI"></a>
+    <a href="https://github.com/silene-systems/catchfly/blob/main/LICENSE"><img src="https://img.shields.io/badge/license-Apache--2.0-green" alt="License"></a>
+    <a href="https://www.python.org/"><img src="https://img.shields.io/pypi/pyversions/catchfly" alt="Python"></a>
+  </p>
+</p>
+
+---
+
+> *Named after [Silene](https://en.wikipedia.org/wiki/Silene) (catchfly) — plants that secrete a sticky substance to capture insects. Catchfly captures scattered terms and groups them into canonical categories.*
+>
+> *Part of the [Silene Systems](https://silene.systems) ecosystem for rare disease research.*
+
+## The problem
+
+You extracted 3,000 mentions from 200 scientific papers. Or an LLM classified 3 million emails into 4,000 labels. Now you have:
+
+- **No schema** — you don't know what categories should exist
+- **Duplicates everywhere** — "miglustat" and "Zavesca" are the same drug
+- **Ambiguous boundaries** — is "cognitive decline" the same as "cognitive impairment"?
+- **Related but distinct entities** — "ALT" and "AST" are similar but clinically different
+
+No existing tool does both **schema discovery** and **term normalization** as a composable Python library.
+
+## What catchfly does
+
+**Two operations, one pipeline:**
+
+**Discover** — *"What categories exist in my data?"*
+
+```python
+from catchfly import Resolver
+
+resolver = Resolver(embed_provider="gemini", llm_provider="openai/gpt-5.4")
+
+schema = resolver.discover(
+    mentions=["miglustat", "splenomegaly", "ALT: 120", "ataxia", ...],
+    contexts={"miglustat": ["Patient received miglustat 200mg daily"], ...},
+)
+
+# schema.categories → ["Treatments", "Symptoms", "Lab Values", ...]
+# schema.examples  → {"Treatments": ["miglustat", "arimoclomol"], ...}
+
+schema.rename("Lab Values", "Laboratory Findings")
+schema.merge("Symptoms", "Clinical Signs")
+```
+
+**Normalize** — *"Which terms belong where, and which are synonyms?"*
+
+```python
+result = resolver.normalize(
+    mentions=all_mentions,
+    schema=schema,
+    contexts=contexts,
+)
+
+# result.groups    → [NormGroup(canonical="miglustat", members=["Zavesca", "NB-DNJ"])]
+# result.ambiguous → [AmbiguousPair("cognitive decline", "cognitive impairment")]
+```
+
+**Or all-in-one:**
+
+```python
+result = resolver.resolve(mentions, contexts=contexts)
+```
+
+## Self-improving prompts
+
+Give catchfly 20–50 labeled examples and it optimizes its own prompts for your domain. Inspired by [GEPA](https://github.com/gepa-ai/gepa), implemented from scratch with zero external dependencies.
+
+```python
+optimized = resolver.optimize(
+    mentions=mentions,
+    contexts=contexts,
+    ground_truth={"miglustat": "miglustat", "Zavesca": "miglustat", "ALT": "ALT", "AST": "AST"},
+    iterations=20,  # ~$1–2, ~15 min
+)
+
+result = optimized.resolve(new_mentions)  # 20–30% better accuracy on your domain
+```
+
+## How it works
+
+Four components, composable as a pipeline:
+
+| Component | What it does | Cost |
+|---|---|---|
+| **EmbeddingSimilarity** | Pre-filter: finds candidate pairs via cosine similarity | Embedding API only |
+| **LLMGrouper** | Core engine: LLM analyzes clusters, proposes categories/synonyms with 4-way relation typing (synonym / hierarchy / related / distinct) | LLM API |
+| **LLMClassifier** | Propagation: assigns remaining mentions to approved categories via few-shot classification | LLM API |
+| **UserSeeded** | User guidance: seed mappings override and anchor discovery + normalization | Embedding only |
+
+## Tiered quality
+
+| Tier | What runs | Cost / 1K mentions | Accuracy | Best for |
+|---|---|---|---|---|
+| **Tier 1** (free) | Embedding only | $0 (local models) | 60–70% | Exploration |
+| **Tier 2** (standard) | Embedding + LLM | $0.15–0.50 | 82–90% | Production |
+| **Tier 3** (optimized) | Tier 2 + optimize() | $1–3 one-time | 90–95%+ | Systematic reviews |
+
+## Use cases
+
+- **Medical literature** — normalize symptoms, drugs, genetic variants from systematic reviews
+- **Email categorization** — collapse 4,000 LLM-generated labels into 150 clean groups
+- **E-commerce tags** — build taxonomy from 50,000 user-generated product tags
+- **Any domain** — if you have messy text labels, catchfly cleans them up
+
+## Provider support
+
+```python
+# Embeddings
+Resolver(embed_provider="gemini")       # Google Gemini Embedding 2
+Resolver(embed_provider="openai")        # OpenAI text-embedding-3
+Resolver(embed_provider="local")         # sentence-transformers (offline, free)
+Resolver(embed_provider=my_function)     # any callable
+
+# LLM
+Resolver(llm_provider="openai/gpt-5.4")
+Resolver(llm_provider="gemini/flash")
+Resolver(llm_provider=my_function)       # any callable
+```
+
+## Evaluation
+
+```python
+ground_truth = {"miglustat": "miglustat", "Zavesca": "miglustat", "ALT": "ALT", "AST": "AST"}
+metrics = resolver.evaluate(result, ground_truth)
+# → precision, recall, false_merges, missed_merges, category_accuracy
+```
+
+## Installation
+
+```bash
+pip install catchfly
+```
+
+> **v0.0.1 is a name reservation.** Active development underway. First functional release (v0.1.0) expected Q2 2026.
+
+## Part of the Silene ecosystem
+
+| Project | What it is |
+|---|---|
+| **[Silene Systems](https://silene.systems)** | Computational phenotyping platform for rare diseases |
+| **Campion** | Agentic batch literature extraction platform (SaaS) |
+| **catchfly** | Schema discovery + term normalization library (open source, this repo) |
+
+> *[Silene tomentosa](https://en.wikipedia.org/wiki/Silene_tomentosa) (Gibraltar campion) is one of the rarest plants in the world — thought extinct in 1992, rediscovered in 1994 on the Rock of Gibraltar. Like rare diseases, it hides in plain sight, waiting for someone to look carefully enough.*
+
+## License
+
+Apache-2.0
+
+## Citation
+
+If you use catchfly in academic work:
+
+```bibtex
+@software{catchfly2026,
+  author = {Michalski, Adrian},
+  title = {catchfly: Schema discovery and category normalization for unstructured text data},
+  year = {2026},
+  url = {https://github.com/silene-systems/catchfly},
+  license = {Apache-2.0}
+}
+``

catchfly-0.0.1/pyproject.toml

@@ -0,0 +1,43 @@
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[project]
+name = "catchfly"
+version = "0.0.1"
+description = "From chaos to categories. Schema discovery + term normalization for unstructured text data."
+readme = "README.md"
+license = "Apache-2.0"
+requires-python = ">=3.10"
+authors = [
+    { name = "Adrian Michalski", email = "adrian@silene.systems" },
+]
+keywords = [
+    "normalization",
+    "schema-discovery",
+    "clustering",
+    "llm",
+    "nlp",
+    "deduplication",
+    "text-classification",
+    "rare-diseases",
+]
+classifiers = [
+    "Development Status :: 1 - Planning",
+    "Intended Audience :: Science/Research",
+    "Intended Audience :: Developers",
+    "License :: OSI Approved :: Apache Software License",
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3.10",
+    "Programming Language :: Python :: 3.11",
+    "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
+    "Topic :: Scientific/Engineering :: Artificial Intelligence",
+    "Topic :: Text Processing :: Linguistic",
+]
+
+[project.urls]
+Homepage = "https://github.com/silene-systems/catchfly"
+Documentation = "https://github.com/silene-systems/catchfly"
+Repository = "https://github.com/silene-systems/catchfly"
+Issues = "https://github.com/silene-systems/catchfly/issues"

catchfly-0.0.1/src/catchfly/__init__.py

@@ -0,0 +1,8 @@
+"""
+catchfly
+
+Schema discovery + category normalization for unstructured text data.
+https://github.com/silene-systems/catchfly
+"""
+
+__version__ = "0.0.1"