relvector 0.1.0__tar.gz

relvector-0.1.0/.gitignore

@@ -0,0 +1,40 @@
+# Python
+__pycache__/
+*.py[cod]
+*.egg-info/
+build/
+dist/
+.eggs/
+# Environments
+.venv/
+.venvs/
+venv/
+# Tooling caches
+.pytest_cache/
+.mypy_cache/
+.ruff_cache/
+.hypothesis/
+.coverage
+.coverage.*
+coverage.xml
+htmlcov/
+# Local artefacts
+/results/
+*.sqlite3
+*.db
+.pgdata/
+.env
+.env.*
+!.env.example
+audit*.jsonl
+.pytest_tmp/
+
+# Static demo build output and download cache
+/.cache/
+/site/
+
+# RelVector local indexes and example output
+*.rvindex
+*.rvindex-*
+relvector_index.sqlite*
+examples/relvector/_output/

relvector-0.1.0/CHANGELOG.md

@@ -0,0 +1,25 @@
+# Changelog
+
+All notable changes are listed here. The project follows semantic versioning once released.
+
+## 0.1.0 (2026-10-25)
+
+First public release:
+
+* Read-only connections for SQLite (`mode=ro`) and PostgreSQL (read-only `REPEATABLE READ`
+  transactions).
+* Schema inspection: tables, columns, primary keys and foreign keys.
+* Entity mappings with one-hop related tables, explicit `Include(via=...)` joins,
+  sensitive-field blocking, exclusions and deterministic templates.
+* Bounded keyset extraction with per-relation `IN` queries and `ROW_NUMBER()` caps that report
+  truncation.
+* Embedding providers: local sentence-transformers and dependency-free hashing.
+* A persistent SQLite vector store with exact cosine search and metadata filters.
+* Idempotent re-indexing with content hashes, pruning of deleted rows, atomic per-batch
+  checkpoints, typed checkpoint keys (date, datetime, UUID, Decimal), resume, and single-writer
+  ownership (`RunSupersededError`).
+* Optional SQLGuard Ninja validation of extraction queries.
+* Streaming exact search: chunked cursor reads, a bounded top-k heap, SQL-side filtering and
+  deterministic tie-breaking. Search memory no longer grows with index size
+  (`LocalVectorStore(search_chunk_size=...)`).
+* PostgreSQL validation tests on a disposable database.

relvector-0.1.0/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.

relvector-0.1.0/PKG-INFO

@@ -0,0 +1,189 @@
+Metadata-Version: 2.5
+Name: relvector
+Version: 0.1.0
+Summary: Schema-aware semantic indexes over existing relational databases, with source references back to the rows.
+Project-URL: Homepage, https://github.com/Pratik653/RelVector
+Project-URL: Issues, https://github.com/Pratik653/RelVector/issues
+Project-URL: Changelog, https://github.com/Pratik653/RelVector/blob/main/CHANGELOG.md
+Author: RelVector contributors
+License-Expression: Apache-2.0
+License-File: LICENSE
+License-File: THIRD_PARTY_NOTICES.md
+Keywords: database,embeddings,rag,semantic-search,sql,sqlalchemy,vector-index
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Developers
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Database
+Classifier: Typing :: Typed
+Requires-Python: >=3.11
+Requires-Dist: numpy>=1.26
+Requires-Dist: sqlalchemy<3.0,>=2.0
+Provides-Extra: dev
+Requires-Dist: build>=1.2; extra == 'dev'
+Requires-Dist: mypy>=1.10; extra == 'dev'
+Requires-Dist: psycopg[binary]>=3.1; extra == 'dev'
+Requires-Dist: pytest>=8.0; extra == 'dev'
+Requires-Dist: ruff>=0.6; extra == 'dev'
+Requires-Dist: sqlguard-ninja<0.8,>=0.7; extra == 'dev'
+Provides-Extra: local
+Requires-Dist: sentence-transformers>=2.3; extra == 'local'
+Provides-Extra: postgres
+Requires-Dist: psycopg[binary]>=3.1; extra == 'postgres'
+Provides-Extra: sqlguard
+Requires-Dist: sqlguard-ninja<0.8,>=0.7; extra == 'sqlguard'
+Description-Content-Type: text/markdown
+
+# RelVector
+
+**Schema-aware semantic indexes over existing relational databases.**
+
+RelVector makes selected relational data semantically searchable without migrating or
+modifying the source database. It reads your schema, follows declared foreign keys, turns each
+root row plus its related rows into deterministic, readable text, embeds it with a
+pluggable provider and stores the vectors in a persistent local index. Every search result
+points back to the exact source rows.
+
+RelVector does **not** convert a database into a vector database. The database stays the
+authoritative source and is only ever read.
+
+> Status: `0.1.0`, alpha. Tested with SQLite and PostgreSQL 16. APIs may change before 1.0.
+
+## Install
+
+Python 3.11+:
+
+```bash
+pip install relvector                # core: SQLAlchemy + NumPy
+pip install "relvector[local]"       # + sentence-transformers for local embeddings (pulls in PyTorch)
+pip install "relvector[postgres]"    # + psycopg 3; use postgresql+psycopg://user@host/db URLs
+pip install "relvector[sqlguard]"    # + optional SQLGuard Ninja query validation
+```
+
+For development, use `pip install -e ".[dev]"` from a checkout.
+
+## Quick start
+
+```python
+from relvector import RelVector
+
+with RelVector("sqlite:///shop.db", embedding_provider="local",
+               index_path="shop.rvindex") as rv:
+    print(rv.inspect_schema().describe())        # tables, columns, PKs, foreign keys
+
+    rv.configure(
+        root_table="orders",                     # one semantic record per order
+        include_tables=["shipments"],            # joined via the declared foreign key
+        text_fields=["orders.status", "shipments.delay_reason"],   # embedded
+        metadata_fields=["orders.status"],       # stored and filterable, not embedded
+        exclude_fields=["customers.email"],
+        templates={
+            "orders": "Order {id} is {status}.",
+            "shipments": "Its associated shipment has a {delay_reason}.",
+        },
+    )
+    stats = rv.index(batch_size=500)
+    print(stats)   # rows, embedded, unchanged, deleted, timings, rows/s
+
+    for hit in rv.search("orders delayed because of transport problems", top_k=5):
+        print(hit.score, hit.source_table, hit.source_pk, hit.references)
+        print(hit.text)
+```
+
+The record for order 1002 reads:
+`Order 1002 is delayed. Its associated shipment has a carrier capacity shortage.`
+
+Runnable example: [examples/relvector/orders_shipments.py](examples/relvector/orders_shipments.py)
+(`--provider hashing` runs without any model).
+
+## How it works
+
+| Stage | Behaviour |
+|---|---|
+| Inspect | SQLAlchemy inspector: tables, columns, primary keys, declared foreign keys. Catalog only, no rows. |
+| Map | `EntityMapping`: root table, directly related tables (FK or explicit `Include(via=...)`), `text_fields`, `metadata_fields`, `exclude_fields`. Only listed columns are ever selected. |
+| Extract | Keyset pagination on the primary key (`WHERE pk > :last ORDER BY pk LIMIT n`). Related rows are fetched per batch with `IN (...)`, one query per relation, so there are no joins and no Cartesian products. One-to-many relations are capped per parent with `ROW_NUMBER()`, the true count is reported in the text and in `truncated`. |
+| Serialise | Deterministic templates. Nulls render as "not recorded"; if a templated field is null, the record falls back to explicit `Field: value.` text instead of producing a misleading sentence. No LLM is involved. |
+| Embed | `"local"` = sentence-transformers (default `all-MiniLM-L6-v2`), `"hashing"` = dependency-free lexical hashing, or any object implementing `EmbeddingProvider`. |
+| Store | `LocalVectorStore`: one SQLite file with float32 vectors, JSON metadata and checkpoints. No pickle. Exact cosine search, streamed in bounded chunks (memory does not grow with the index). |
+| Re-index | Idempotent: vector ids are derived from the primary key, unchanged content (by hash) is not re-embedded, rows that disappeared are pruned when a run completes. Each batch commits atomically with its checkpoint; `index(resume=True)` continues an interrupted run. |
+
+Search results carry `vector_id`, `score`, `text`, `source_table`, `source_pk`, `metadata`,
+`references` (the related rows used, by primary key) and `truncated`. Metadata can be filtered
+exactly: `rv.search(q, filters={"orders.status": "delayed"})`.
+
+**A high similarity score is not evidence of causation or of a fact.** It means the text of
+a record is close to the query in embedding space. Check the source rows.
+
+## Safety defaults
+
+* Source connections are read-only: SQLite opens with `mode=ro` (a missing file is an error,
+  never created); PostgreSQL sessions use `READ ONLY`, `REPEATABLE READ` transactions. Use a
+  read-only database role as well. If you pass your own `engine=`, these settings are yours to
+  configure.
+* Only `SELECT` statements built with SQLAlchemy Core run, with bound parameters. There is no
+  free-form SQL input.
+* Columns whose names look sensitive (email, phone, password, SSN, card number and so on) are
+  refused in `text_fields`/`metadata_fields` unless listed in `allow_sensitive`. Name heuristics
+  are incomplete: `notes` may hold anything, so review every field you map.
+* Providers that send text off the machine must declare `sends_data_externally = True` and
+  are refused unless you pass `allow_external_embeddings=True`.
+* `repr()` hides passwords in connection URLs.
+* Optional [SQLGuard Ninja](https://pypi.org/project/sqlguard-ninja/) integration validates every extraction
+  query against a policy and blocks on denial:
+
+  ```python
+  from sqlguard_ninja import Guard, Policy
+  from relvector.integrations.sqlguard import SQLGuardValidator
+
+  rv = RelVector(url, sql_validator=SQLGuardValidator(Guard(Policy.from_file("policy.yaml")),
+                                                      role="indexer"))
+  ```
+
+  SQLGuard approval means the query is allowed for the role. It does not mean the values may
+  be embedded and stored; that decision belongs to the mapping.
+
+## Known limitations
+
+* This is a trusted-developer library. The index file contains embedded text and metadata in
+  plain form. Anyone who can read it can read that data. There is no row-level or tenant
+  isolation in the index.
+* Relationships: one hop from the root table, single-column joins. No self-references,
+  multi-hop paths or composite foreign keys yet. Composite root primary keys are supported.
+* Re-indexing is a full re-scan. There is no change-data-capture, and no triggers or timestamps
+  are used. Edits to related rows (child shipments, a parent customer) are picked up on the
+  next completed run. Rows changed during a run may be read at different moments, and the next
+  run corrects them. See
+  [how source changes reach documents](docs/relvector/architecture.md#how-source-changes-reach-semantic-documents).
+* Only one writer per entity at a time. A new run, or a resume of the same run in another
+  process, takes over the checkpoint. The displaced writer stops with `RunSupersededError` at
+  its next batch and never prunes. This is a takeover, not a lock: it doesn't stop two writers
+  from embedding the same batch, and those duplicate writes are harmless upserts.
+* Checkpoints keep primary-key types (int, str, date, datetime, time, Decimal, UUID, bytes), so
+  a resumed scan compares keys as the column does. Other key types cannot be checkpointed and
+  fail with a clear error.
+* `LocalVectorStore` search is exact and streamed: memory is bounded by
+  `search_chunk_size` (default 1024 rows) and `top_k`, but every query scans all vectors that
+  pass the filters. Latency grows linearly with index size, and there is no approximate (ANN)
+  index. See [search performance](docs/relvector/architecture.md#search-performance).
+* The hashing provider is lexical, not semantic.
+* SQLite and PostgreSQL only. Other databases are not supported.
+
+## Development
+
+```bash
+pip install -e ".[dev]"
+pytest                      # model and PostgreSQL tests skip unless available (see CONTRIBUTING.md)
+ruff check src/relvector tests/relvector examples/relvector
+mypy
+```
+
+Architecture, reuse decisions and roadmap: [docs/relvector/architecture.md](docs/relvector/architecture.md).
+
+## License
+
+Apache-2.0. See [LICENSE](LICENSE) and [THIRD_PARTY_NOTICES.md](THIRD_PARTY_NOTICES.md). This
+repository grew out of SQLGuard Ninja (same authors and licence), whose history it keeps.

relvector-0.1.0/README.md

@@ -0,0 +1,150 @@
+# RelVector
+
+**Schema-aware semantic indexes over existing relational databases.**
+
+RelVector makes selected relational data semantically searchable without migrating or
+modifying the source database. It reads your schema, follows declared foreign keys, turns each
+root row plus its related rows into deterministic, readable text, embeds it with a
+pluggable provider and stores the vectors in a persistent local index. Every search result
+points back to the exact source rows.
+
+RelVector does **not** convert a database into a vector database. The database stays the
+authoritative source and is only ever read.
+
+> Status: `0.1.0`, alpha. Tested with SQLite and PostgreSQL 16. APIs may change before 1.0.
+
+## Install
+
+Python 3.11+:
+
+```bash
+pip install relvector                # core: SQLAlchemy + NumPy
+pip install "relvector[local]"       # + sentence-transformers for local embeddings (pulls in PyTorch)
+pip install "relvector[postgres]"    # + psycopg 3; use postgresql+psycopg://user@host/db URLs
+pip install "relvector[sqlguard]"    # + optional SQLGuard Ninja query validation
+```
+
+For development, use `pip install -e ".[dev]"` from a checkout.
+
+## Quick start
+
+```python
+from relvector import RelVector
+
+with RelVector("sqlite:///shop.db", embedding_provider="local",
+               index_path="shop.rvindex") as rv:
+    print(rv.inspect_schema().describe())        # tables, columns, PKs, foreign keys
+
+    rv.configure(
+        root_table="orders",                     # one semantic record per order
+        include_tables=["shipments"],            # joined via the declared foreign key
+        text_fields=["orders.status", "shipments.delay_reason"],   # embedded
+        metadata_fields=["orders.status"],       # stored and filterable, not embedded
+        exclude_fields=["customers.email"],
+        templates={
+            "orders": "Order {id} is {status}.",
+            "shipments": "Its associated shipment has a {delay_reason}.",
+        },
+    )
+    stats = rv.index(batch_size=500)
+    print(stats)   # rows, embedded, unchanged, deleted, timings, rows/s
+
+    for hit in rv.search("orders delayed because of transport problems", top_k=5):
+        print(hit.score, hit.source_table, hit.source_pk, hit.references)
+        print(hit.text)
+```
+
+The record for order 1002 reads:
+`Order 1002 is delayed. Its associated shipment has a carrier capacity shortage.`
+
+Runnable example: [examples/relvector/orders_shipments.py](examples/relvector/orders_shipments.py)
+(`--provider hashing` runs without any model).
+
+## How it works
+
+| Stage | Behaviour |
+|---|---|
+| Inspect | SQLAlchemy inspector: tables, columns, primary keys, declared foreign keys. Catalog only, no rows. |
+| Map | `EntityMapping`: root table, directly related tables (FK or explicit `Include(via=...)`), `text_fields`, `metadata_fields`, `exclude_fields`. Only listed columns are ever selected. |
+| Extract | Keyset pagination on the primary key (`WHERE pk > :last ORDER BY pk LIMIT n`). Related rows are fetched per batch with `IN (...)`, one query per relation, so there are no joins and no Cartesian products. One-to-many relations are capped per parent with `ROW_NUMBER()`, the true count is reported in the text and in `truncated`. |
+| Serialise | Deterministic templates. Nulls render as "not recorded"; if a templated field is null, the record falls back to explicit `Field: value.` text instead of producing a misleading sentence. No LLM is involved. |
+| Embed | `"local"` = sentence-transformers (default `all-MiniLM-L6-v2`), `"hashing"` = dependency-free lexical hashing, or any object implementing `EmbeddingProvider`. |
+| Store | `LocalVectorStore`: one SQLite file with float32 vectors, JSON metadata and checkpoints. No pickle. Exact cosine search, streamed in bounded chunks (memory does not grow with the index). |
+| Re-index | Idempotent: vector ids are derived from the primary key, unchanged content (by hash) is not re-embedded, rows that disappeared are pruned when a run completes. Each batch commits atomically with its checkpoint; `index(resume=True)` continues an interrupted run. |
+
+Search results carry `vector_id`, `score`, `text`, `source_table`, `source_pk`, `metadata`,
+`references` (the related rows used, by primary key) and `truncated`. Metadata can be filtered
+exactly: `rv.search(q, filters={"orders.status": "delayed"})`.
+
+**A high similarity score is not evidence of causation or of a fact.** It means the text of
+a record is close to the query in embedding space. Check the source rows.
+
+## Safety defaults
+
+* Source connections are read-only: SQLite opens with `mode=ro` (a missing file is an error,
+  never created); PostgreSQL sessions use `READ ONLY`, `REPEATABLE READ` transactions. Use a
+  read-only database role as well. If you pass your own `engine=`, these settings are yours to
+  configure.
+* Only `SELECT` statements built with SQLAlchemy Core run, with bound parameters. There is no
+  free-form SQL input.
+* Columns whose names look sensitive (email, phone, password, SSN, card number and so on) are
+  refused in `text_fields`/`metadata_fields` unless listed in `allow_sensitive`. Name heuristics
+  are incomplete: `notes` may hold anything, so review every field you map.
+* Providers that send text off the machine must declare `sends_data_externally = True` and
+  are refused unless you pass `allow_external_embeddings=True`.
+* `repr()` hides passwords in connection URLs.
+* Optional [SQLGuard Ninja](https://pypi.org/project/sqlguard-ninja/) integration validates every extraction
+  query against a policy and blocks on denial:
+
+  ```python
+  from sqlguard_ninja import Guard, Policy
+  from relvector.integrations.sqlguard import SQLGuardValidator
+
+  rv = RelVector(url, sql_validator=SQLGuardValidator(Guard(Policy.from_file("policy.yaml")),
+                                                      role="indexer"))
+  ```
+
+  SQLGuard approval means the query is allowed for the role. It does not mean the values may
+  be embedded and stored; that decision belongs to the mapping.
+
+## Known limitations
+
+* This is a trusted-developer library. The index file contains embedded text and metadata in
+  plain form. Anyone who can read it can read that data. There is no row-level or tenant
+  isolation in the index.
+* Relationships: one hop from the root table, single-column joins. No self-references,
+  multi-hop paths or composite foreign keys yet. Composite root primary keys are supported.
+* Re-indexing is a full re-scan. There is no change-data-capture, and no triggers or timestamps
+  are used. Edits to related rows (child shipments, a parent customer) are picked up on the
+  next completed run. Rows changed during a run may be read at different moments, and the next
+  run corrects them. See
+  [how source changes reach documents](docs/relvector/architecture.md#how-source-changes-reach-semantic-documents).
+* Only one writer per entity at a time. A new run, or a resume of the same run in another
+  process, takes over the checkpoint. The displaced writer stops with `RunSupersededError` at
+  its next batch and never prunes. This is a takeover, not a lock: it doesn't stop two writers
+  from embedding the same batch, and those duplicate writes are harmless upserts.
+* Checkpoints keep primary-key types (int, str, date, datetime, time, Decimal, UUID, bytes), so
+  a resumed scan compares keys as the column does. Other key types cannot be checkpointed and
+  fail with a clear error.
+* `LocalVectorStore` search is exact and streamed: memory is bounded by
+  `search_chunk_size` (default 1024 rows) and `top_k`, but every query scans all vectors that
+  pass the filters. Latency grows linearly with index size, and there is no approximate (ANN)
+  index. See [search performance](docs/relvector/architecture.md#search-performance).
+* The hashing provider is lexical, not semantic.
+* SQLite and PostgreSQL only. Other databases are not supported.
+
+## Development
+
+```bash
+pip install -e ".[dev]"
+pytest                      # model and PostgreSQL tests skip unless available (see CONTRIBUTING.md)
+ruff check src/relvector tests/relvector examples/relvector
+mypy
+```
+
+Architecture, reuse decisions and roadmap: [docs/relvector/architecture.md](docs/relvector/architecture.md).
+
+## License
+
+Apache-2.0. See [LICENSE](LICENSE) and [THIRD_PARTY_NOTICES.md](THIRD_PARTY_NOTICES.md). This
+repository grew out of SQLGuard Ninja (same authors and licence), whose history it keeps.

relvector-0.1.0/THIRD_PARTY_NOTICES.md

@@ -0,0 +1,54 @@
+# Third-party notices
+
+## RelVector (the `relvector` wheel)
+
+RelVector is licensed under Apache-2.0. **It does not vendor or copy third-party source code.**
+The one adapted piece is the list of sensitive-column name patterns in
+`src/relvector/mapping.py`. It comes from SQLGuard Ninja (`sqlguard_ninja/privacy/suggest.py`),
+which has the same authors and the same Apache-2.0 licence.
+
+The packages below are ordinary dependencies, installed by the package manager under their
+own licences. Licences were read from installed package metadata on 2026-10-25. Always check
+the version you install.
+
+### Runtime dependencies (core install)
+
+| Package | Version tested | Licence | Use |
+|---|---|---|---|
+| [SQLAlchemy](https://www.sqlalchemy.org/) | 2.1.6 | MIT | Schema inspection, query construction, connections |
+| [NumPy](https://numpy.org/) | 2.4.6 | BSD-3-Clause (bundled parts: 0BSD, MIT, Zlib, CC0-1.0) | Vector arithmetic |
+
+### Optional extras
+
+| Extra | Package | Version tested | Licence |
+|---|---|---|---|
+| `local` | sentence-transformers | 6.2.0 | Apache-2.0 |
+| `local` (transitive) | torch / transformers | 2.14.1 (CPU) / 5.20.2 | BSD-3-Clause and others (see the package) / Apache-2.0 |
+| `postgres` | psycopg (+ psycopg-binary) | 3.3.6 | LGPL-3.0-only |
+| `sqlguard` | sqlguard-ninja | 0.7.0 | Apache-2.0 |
+
+psycopg is LGPL-3.0. RelVector does not import it directly; SQLAlchemy loads it at runtime when
+a `postgresql+psycopg://` URL is used. It is not modified or bundled.
+
+Embedding models are not part of RelVector. The default local model,
+`sentence-transformers/all-MiniLM-L6-v2`, is published under Apache-2.0. Models are downloaded
+by sentence-transformers under their own licences; check the licence of any model you configure.
+
+### Development and test tools (not distributed)
+
+pytest (MIT), ruff (MIT), mypy (MIT), build (MIT), twine (Apache-2.0). PostgreSQL tests can
+use `pgserver` (not a declared dependency; it bundles PostgreSQL binaries under the
+PostgreSQL License) or a PostgreSQL 16 service container in CI.
+
+## Origin of this repository
+
+RelVector's repository began as a copy of SQLGuard Ninja (Apache-2.0, same authors). The
+SQLGuard Ninja sources, benchmarks and their third-party notices have been removed from the
+working tree. They remain in this repository's git history, and SQLGuard Ninja is developed
+and published separately (`pip install sqlguard-ninja`). RelVector uses it only as an optional
+dependency.
+
+## GitHub Actions
+
+`actions/checkout`, `actions/setup-python` and `actions/upload-artifact` (MIT, GitHub) are
+referenced by commit SHA in the workflows.