lede-spacy 0.3.0__tar.gz

lede_spacy-0.3.0/.gitignore

@@ -0,0 +1,53 @@
+# Python
+__pycache__/
+*.py[cod]
+*.egg-info/
+.pytest_cache/
+.coverage
+build/
+dist/
+.venv/
+venv/
+
+# IDE
+.vscode/
+.idea/
+
+# Per-developer agent / tooling config (Claude Code, etc.)
+# These files often contain machine-local paths and per-machine permission
+# allowlists that should not be shared.
+.claude/
+!.claude/agents/
+!.claude/commands/
+
+# Skill output (research/audit artifacts — point-in-time, not source).
+# Per the skill-output-hygiene rule. Never commit these to a public repo.
+skill-output/
+
+# Secrets — defensive, even though none exist today
+.env
+.env.*
+!.env.example
+.openai
+.openai.*
+.anthropic
+.anthropic.*
+*.pem
+*.key
+*.p12
+*.pfx
+*.jks
+id_rsa
+id_ed25519
+credentials*.json
+secrets*.json
+
+# Terraform / IaC state
+*.tfstate
+*.tfstate.*
+*.tfvars
+!*.tfvars.example
+
+# OS cruft
+.DS_Store
+Thumbs.db

lede_spacy-0.3.0/LICENSE

@@ -0,0 +1,202 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

lede_spacy-0.3.0/PKG-INFO

@@ -0,0 +1,226 @@
+Metadata-Version: 2.4
+Name: lede-spacy
+Version: 0.3.0
+Summary: spaCy-powered enrichment backend for lede — PERSON/ORG/GPE entity extraction via en_core_web_sm.
+Project-URL: Homepage, https://github.com/yonk-labs/lede
+Project-URL: Repository, https://github.com/yonk-labs/lede
+Project-URL: Issues, https://github.com/yonk-labs/lede/issues
+Author: Yonk
+License-Expression: Apache-2.0
+License-File: LICENSE
+Keywords: entity-extraction,lede,ner,spacy
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+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 :: Text Processing :: Linguistic
+Requires-Python: >=3.10
+Requires-Dist: lede>=0.3.0
+Requires-Dist: spacy<3.9,>=3.8
+Provides-Extra: dev
+Requires-Dist: pytest>=7; extra == 'dev'
+Description-Content-Type: text/markdown
+
+# lede-spacy
+
+spaCy-powered enrichment backend for [lede](https://github.com/yonk-labs/lede).
+Adds proper named-entity recognition (PERSON / ORG / GPE) and richer
+phrase / fact extraction by registering itself as the `"spacy"` backend
+for `lede.extract.metadata`, `lede.extract.phrases`, and
+`lede.extract.correlate_facts`.
+
+| | regex backend (default in `lede`) | spaCy backend (this package) |
+|---|---|---|
+| Entities (PERSON / ORG / GPE) | always empty | populated from `en_core_web_sm` |
+| Phrases | repeated multi-word n-grams | syntactically-grounded noun chunks (still count-filtered) |
+| Correlate facts | regex pattern → entity↔number | dependency-parse → wider net of entity↔number relationships |
+| Latency | sub-millisecond | ~5 ms after warmup, ~50 ms first call (model load) |
+| Determinism | byte-identical Python ↔ Rust | spaCy is deterministic but Python-only and not byte-comparable to Rust |
+| Install footprint | stdlib only | `spacy>=3.8` + `en_core_web_sm` (~50 MB model) |
+
+## When you actually want this
+
+lede's regex backend covers the majority of structured-extract use cases for
+free — dates, amounts, URLs, numeric facts with sentence context, and
+repeated-phrase mining all work with zero dependencies. **You want
+lede-spacy specifically when you need named entities** — people,
+companies, places — pulled out of arbitrary text. That's where the regex
+backend explicitly returns nothing.
+
+You also get richer `correlate_facts` (the dep-parser walks the syntax
+tree to find entity↔number relationships even for entities mentioned
+once, where the regex backend requires repetition).
+
+## Side-by-side: the same input, both backends
+
+```python
+from lede.extract import metadata
+import lede_spacy  # side effect: registers the 'spacy' backend
+```
+
+The text:
+
+> Acme Corp announced today a partnership with Yonk Labs to integrate
+> deterministic summarization into their RAG pipeline. The deal,
+> brokered by CEO Lin Wu and signed in San Francisco on 2024-11-15,
+> covers $2.4M in annual licensing through 2027. Sarah Jones from
+> Acme's engineering team and Marcus Chen from Yonk Labs will lead
+> the joint integration. The first deployment is targeted for European
+> customers, including teams in London, Berlin, and Paris.
+
+### `backend="regex"` (lede default — zero-dep)
+
+```python
+m = metadata(text)
+m.dates     # ('2024-11-15', '2027')
+m.amounts   # ('$2.4M',)
+m.urls      # ()
+m.entities  # ()    ← regex backend returns nothing here
+```
+
+The regex backend caught the structured stuff (ISO date, year, dollar
+amount). It can't do entities — that's not a regex job.
+
+### `backend="spacy"` (this package)
+
+```python
+m = metadata(text, backend="spacy")
+m.dates     # ('2024-11-15', '2027')        — same regex stage runs
+m.amounts   # ('$2.4M',)                    — same regex stage runs
+m.urls      # ()                            — same regex stage runs
+m.entities  # ('Acme Corp', 'Yonk Labs', 'RAG', 'Lin Wu',
+            #  'San Francisco', 'Sarah Jones', 'Acme',
+            #  'Marcus Chen', 'London', 'Berlin', 'Paris')
+```
+
+11 entities pulled out of the same input. PERSON ('Lin Wu', 'Sarah
+Jones', 'Marcus Chen'), ORG ('Acme Corp', 'Yonk Labs', 'Acme', 'RAG'),
+GPE ('San Francisco', 'London', 'Berlin', 'Paris'). The dates / amounts /
+URLs columns are unchanged — spaCy backend runs the same regex stages
+*plus* the spaCy NER stage on top.
+
+### Use case: `correlate_facts` finds different relationships
+
+For some inputs the two backends produce *different but overlapping* fact
+relationships. From the same paragraph above:
+
+```python
+correlate_facts(text)                        # regex: 2 pairs anchored on Acme Corp
+correlate_facts(text, backend="spacy")       # spaCy: 4 pairs, also catches Yonk Labs and customer churn
+```
+
+The dep-parser approach catches relationships the regex misses,
+especially for entities that appear once. (And vice versa — sometimes the
+regex backend catches a pattern the dep-parser doesn't. The two are
+complementary, not strictly better/worse. Switch backend per call if you
+need it.)
+
+## Use this when…
+
+- Your callers want **PERSON / ORG / GPE entities** in the output. Default lede can't help.
+- You want richer entity-number correlations on documents where each entity is mentioned once.
+- You're already running spaCy in your pipeline and want to consolidate.
+- Latency budgets are in the 5–50 ms range per chunk, not sub-millisecond.
+
+## Don't use this when…
+
+- You need **byte-identical Python ↔ Rust output**. spaCy is Python-only and isn't on the parity contract.
+- You're on a sub-millisecond hot path. spaCy is ~5 ms per call after warmup, ~50 ms first call.
+- You don't actually need entities. The default lede regex backend already handles dates / amounts / URLs / numeric facts with sentence context.
+- You're shipping lede inside a constrained environment (Lambda cold-start, embedded, no-egress) — the 50 MB `en_core_web_sm` model has real cost.
+- You can't tolerate spaCy's transitive dependency graph (NumPy, Cython, blis, thinc, etc.) in your env.
+
+## Install
+
+```bash
+pip install lede-spacy
+python -m spacy download en_core_web_sm
+```
+
+The first command pulls `lede` and `spacy>=3.8,<3.9`. The second pulls
+the ~50 MB `en_core_web_sm` 3.8.0 model. PyPI does not allow direct-URL
+dependencies, so the model is a separate install step (the same
+convention spaCy itself uses).
+
+If you want a single reproducible install, pin the model wheel from
+`requirements.txt`:
+
+```
+lede-spacy==0.3.0
+https://github.com/explosion/spacy-models/releases/download/en_core_web_sm-3.8.0/en_core_web_sm-3.8.0-py3-none-any.whl
+```
+
+From source (in the lede repo):
+
+```bash
+pip install -e packages/lede-spacy
+python -m spacy download en_core_web_sm
+```
+
+## Use
+
+```python
+import lede_spacy            # side-effect: registers the "spacy" backend
+from lede.extract import metadata, phrases, correlate_facts
+
+# Per-call backend override:
+m = metadata(text, backend="spacy")
+
+# Or set the global default once:
+import lede
+lede.set_default_backend("auto")   # spaCy if registered, else regex
+```
+
+Pre-load the model once at startup to avoid the ~50 ms first-call model
+load:
+
+```python
+from lede_spacy import warmup
+warmup()
+```
+
+## Performance
+
+| call | latency (typical) |
+|---|---|
+| First call (cold model) | 50–80 ms |
+| Subsequent calls (warm) | ~5 ms |
+| Default lede regex backend (for comparison) | <1 ms |
+
+Run `from lede_spacy import warmup; warmup()` at app startup to pay the
+50 ms once instead of on the first user request.
+
+## What's registered
+
+When you `import lede_spacy`, three backends register themselves into
+`lede.extract._backends`:
+
+| lede primitive | spaCy backend |
+|---|---|
+| `metadata(text, backend="spacy")` | runs regex `dates`/`amounts`/`urls` + spaCy NER for `entities` |
+| `phrases(text, backend="spacy")` | `doc.noun_chunks` filtered to repeated multi-word chunks (matches the regex backend's count semantics) |
+| `correlate_facts(text, backend="spacy")` | DepMatcher-based entity↔number pairing |
+
+The `regex` backend stays the default — `import lede_spacy` is purely
+additive. Existing callers using `backend="regex"` (or no `backend=`
+kwarg) see no behavior change.
+
+## Determinism + parity
+
+spaCy is deterministic per-version: `en_core_web_sm` 3.8.0 produces the
+same entities for the same input on every call. **It is not on lede's
+Python ↔ Rust parity contract**, by design. The Rust port has no spaCy
+equivalent and `Metadata.entities` stays empty under any Rust call. See
+[`docs/lede-spacy-integration.md`](https://github.com/yonk-labs/lede/blob/main/docs/lede-spacy-integration.md)
+for the cross-language policy.
+
+If you need NER from a Rust service today: call out to a Python
+lede-spacy worker, or to a hosted NER endpoint. A future
+`lede-rust-ner` companion crate is on the roadmap if there's demand —
+file an issue.
+
+## License
+
+Apache-2.0, same as lede.

lede_spacy-0.3.0/README.md

@@ -0,0 +1,201 @@
+# lede-spacy
+
+spaCy-powered enrichment backend for [lede](https://github.com/yonk-labs/lede).
+Adds proper named-entity recognition (PERSON / ORG / GPE) and richer
+phrase / fact extraction by registering itself as the `"spacy"` backend
+for `lede.extract.metadata`, `lede.extract.phrases`, and
+`lede.extract.correlate_facts`.
+
+| | regex backend (default in `lede`) | spaCy backend (this package) |
+|---|---|---|
+| Entities (PERSON / ORG / GPE) | always empty | populated from `en_core_web_sm` |
+| Phrases | repeated multi-word n-grams | syntactically-grounded noun chunks (still count-filtered) |
+| Correlate facts | regex pattern → entity↔number | dependency-parse → wider net of entity↔number relationships |
+| Latency | sub-millisecond | ~5 ms after warmup, ~50 ms first call (model load) |
+| Determinism | byte-identical Python ↔ Rust | spaCy is deterministic but Python-only and not byte-comparable to Rust |
+| Install footprint | stdlib only | `spacy>=3.8` + `en_core_web_sm` (~50 MB model) |
+
+## When you actually want this
+
+lede's regex backend covers the majority of structured-extract use cases for
+free — dates, amounts, URLs, numeric facts with sentence context, and
+repeated-phrase mining all work with zero dependencies. **You want
+lede-spacy specifically when you need named entities** — people,
+companies, places — pulled out of arbitrary text. That's where the regex
+backend explicitly returns nothing.
+
+You also get richer `correlate_facts` (the dep-parser walks the syntax
+tree to find entity↔number relationships even for entities mentioned
+once, where the regex backend requires repetition).
+
+## Side-by-side: the same input, both backends
+
+```python
+from lede.extract import metadata
+import lede_spacy  # side effect: registers the 'spacy' backend
+```
+
+The text:
+
+> Acme Corp announced today a partnership with Yonk Labs to integrate
+> deterministic summarization into their RAG pipeline. The deal,
+> brokered by CEO Lin Wu and signed in San Francisco on 2024-11-15,
+> covers $2.4M in annual licensing through 2027. Sarah Jones from
+> Acme's engineering team and Marcus Chen from Yonk Labs will lead
+> the joint integration. The first deployment is targeted for European
+> customers, including teams in London, Berlin, and Paris.
+
+### `backend="regex"` (lede default — zero-dep)
+
+```python
+m = metadata(text)
+m.dates     # ('2024-11-15', '2027')
+m.amounts   # ('$2.4M',)
+m.urls      # ()
+m.entities  # ()    ← regex backend returns nothing here
+```
+
+The regex backend caught the structured stuff (ISO date, year, dollar
+amount). It can't do entities — that's not a regex job.
+
+### `backend="spacy"` (this package)
+
+```python
+m = metadata(text, backend="spacy")
+m.dates     # ('2024-11-15', '2027')        — same regex stage runs
+m.amounts   # ('$2.4M',)                    — same regex stage runs
+m.urls      # ()                            — same regex stage runs
+m.entities  # ('Acme Corp', 'Yonk Labs', 'RAG', 'Lin Wu',
+            #  'San Francisco', 'Sarah Jones', 'Acme',
+            #  'Marcus Chen', 'London', 'Berlin', 'Paris')
+```
+
+11 entities pulled out of the same input. PERSON ('Lin Wu', 'Sarah
+Jones', 'Marcus Chen'), ORG ('Acme Corp', 'Yonk Labs', 'Acme', 'RAG'),
+GPE ('San Francisco', 'London', 'Berlin', 'Paris'). The dates / amounts /
+URLs columns are unchanged — spaCy backend runs the same regex stages
+*plus* the spaCy NER stage on top.
+
+### Use case: `correlate_facts` finds different relationships
+
+For some inputs the two backends produce *different but overlapping* fact
+relationships. From the same paragraph above:
+
+```python
+correlate_facts(text)                        # regex: 2 pairs anchored on Acme Corp
+correlate_facts(text, backend="spacy")       # spaCy: 4 pairs, also catches Yonk Labs and customer churn
+```
+
+The dep-parser approach catches relationships the regex misses,
+especially for entities that appear once. (And vice versa — sometimes the
+regex backend catches a pattern the dep-parser doesn't. The two are
+complementary, not strictly better/worse. Switch backend per call if you
+need it.)
+
+## Use this when…
+
+- Your callers want **PERSON / ORG / GPE entities** in the output. Default lede can't help.
+- You want richer entity-number correlations on documents where each entity is mentioned once.
+- You're already running spaCy in your pipeline and want to consolidate.
+- Latency budgets are in the 5–50 ms range per chunk, not sub-millisecond.
+
+## Don't use this when…
+
+- You need **byte-identical Python ↔ Rust output**. spaCy is Python-only and isn't on the parity contract.
+- You're on a sub-millisecond hot path. spaCy is ~5 ms per call after warmup, ~50 ms first call.
+- You don't actually need entities. The default lede regex backend already handles dates / amounts / URLs / numeric facts with sentence context.
+- You're shipping lede inside a constrained environment (Lambda cold-start, embedded, no-egress) — the 50 MB `en_core_web_sm` model has real cost.
+- You can't tolerate spaCy's transitive dependency graph (NumPy, Cython, blis, thinc, etc.) in your env.
+
+## Install
+
+```bash
+pip install lede-spacy
+python -m spacy download en_core_web_sm
+```
+
+The first command pulls `lede` and `spacy>=3.8,<3.9`. The second pulls
+the ~50 MB `en_core_web_sm` 3.8.0 model. PyPI does not allow direct-URL
+dependencies, so the model is a separate install step (the same
+convention spaCy itself uses).
+
+If you want a single reproducible install, pin the model wheel from
+`requirements.txt`:
+
+```
+lede-spacy==0.3.0
+https://github.com/explosion/spacy-models/releases/download/en_core_web_sm-3.8.0/en_core_web_sm-3.8.0-py3-none-any.whl
+```
+
+From source (in the lede repo):
+
+```bash
+pip install -e packages/lede-spacy
+python -m spacy download en_core_web_sm
+```
+
+## Use
+
+```python
+import lede_spacy            # side-effect: registers the "spacy" backend
+from lede.extract import metadata, phrases, correlate_facts
+
+# Per-call backend override:
+m = metadata(text, backend="spacy")
+
+# Or set the global default once:
+import lede
+lede.set_default_backend("auto")   # spaCy if registered, else regex
+```
+
+Pre-load the model once at startup to avoid the ~50 ms first-call model
+load:
+
+```python
+from lede_spacy import warmup
+warmup()
+```
+
+## Performance
+
+| call | latency (typical) |
+|---|---|
+| First call (cold model) | 50–80 ms |
+| Subsequent calls (warm) | ~5 ms |
+| Default lede regex backend (for comparison) | <1 ms |
+
+Run `from lede_spacy import warmup; warmup()` at app startup to pay the
+50 ms once instead of on the first user request.
+
+## What's registered
+
+When you `import lede_spacy`, three backends register themselves into
+`lede.extract._backends`:
+
+| lede primitive | spaCy backend |
+|---|---|
+| `metadata(text, backend="spacy")` | runs regex `dates`/`amounts`/`urls` + spaCy NER for `entities` |
+| `phrases(text, backend="spacy")` | `doc.noun_chunks` filtered to repeated multi-word chunks (matches the regex backend's count semantics) |
+| `correlate_facts(text, backend="spacy")` | DepMatcher-based entity↔number pairing |
+
+The `regex` backend stays the default — `import lede_spacy` is purely
+additive. Existing callers using `backend="regex"` (or no `backend=`
+kwarg) see no behavior change.
+
+## Determinism + parity
+
+spaCy is deterministic per-version: `en_core_web_sm` 3.8.0 produces the
+same entities for the same input on every call. **It is not on lede's
+Python ↔ Rust parity contract**, by design. The Rust port has no spaCy
+equivalent and `Metadata.entities` stays empty under any Rust call. See
+[`docs/lede-spacy-integration.md`](https://github.com/yonk-labs/lede/blob/main/docs/lede-spacy-integration.md)
+for the cross-language policy.
+
+If you need NER from a Rust service today: call out to a Python
+lede-spacy worker, or to a hosted NER endpoint. A future
+`lede-rust-ner` companion crate is on the roadmap if there's demand —
+file an issue.
+
+## License
+
+Apache-2.0, same as lede.