HeOCR · shaypal5 · May 25, 2026 · May 25, 2026
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
@@ -29,7 +29,7 @@ jobs:
       - name: Check out upstream scans repo for cross-validation
         uses: actions/checkout@v4
         with:
-          repository: HeOCR/public-domain-hand-written-hebrew-scans
+          repository: HeOCR/hash
           path: .upstream
           lfs: false
       - name: Install dev dependencies

diff --git a/AGENTS.md b/AGENTS.md
@@ -11,7 +11,7 @@ policy.
 A dataset of **sets of per-letter images of handwritten Hebrew letters**,
 grouped by writer. Each set = one person/scribe. Each per-letter image
 is a **crop** of a permissively-licensed upstream scan from
-[HeOCR/public-domain-hand-written-hebrew-scans][upstream], with rights
+[HeOCR/hash][upstream], with rights
 inherited and recorded per image. Canonical layout, schema motivation,
 and ingestion model live in [`docs/dataset_structure.md`]\
 (docs/dataset_structure.md). The Hebrew letter enumeration is in
@@ -22,7 +22,7 @@ per-image rights inheritance) is described in
 [`schemas/entry.schema.json`](schemas/entry.schema.json). The release
 runbook is [`docs/release_process.md`](docs/release_process.md).
 
-[upstream]: https://github.com/HeOCR/public-domain-hand-written-hebrew-scans
+[upstream]: https://github.com/HeOCR/hash
 
 ## First-time setup
 
@@ -72,7 +72,7 @@ against the live upstream entry records:
 
 ```bash
 python3 scripts/validate_indexes.py \
-  --upstream-path ../public-domain-hand-written-hebrew-scans
+  --upstream-path ../hash
 ```
 
 CI checks out the upstream repo as a sibling and runs the validator with

diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -20,7 +20,7 @@ licensing policy needed to start ingesting.
 
 - Writer-level (`schemas/writer.schema.json`) and entry-level
   (`schemas/entry.schema.json`) record contracts. Each entry references
-  an upstream scan in `HeOCR/public-domain-hand-written-hebrew-scans`
+  an upstream scan in `HeOCR/hash`
   by `source_id`, `entry_id`, `sha256` (mutable-tag-free), `commit`
   (40-char SHA), and `bbox`.
 - `scripts/validate_indexes.py`: schema validation, referential

diff --git a/CITATION.cff b/CITATION.cff
@@ -3,7 +3,7 @@ cff-version: 1.2.0
 message: Please cite this dataset using the metadata below.
 type: dataset
 title: Hebrew Handwritten Per-Letter Image Dataset
-abstract: Per-letter image crops of handwritten Hebrew letters, grouped into sets by writer. Each crop is a derivative of a permissively-licensed upstream scan in HeOCR/public-domain-hand-written-hebrew-scans, with per-image rights inherited and attribution recorded. The index is line-oriented JSON (JSONL). Release 0.0.0-rc contains 48 per-letter image entries drawn from 2 verified writers (18 LicenseRef-Public-Domain-Israel, 30 PDM-1.0).
+abstract: Per-letter image crops of handwritten Hebrew letters, grouped into sets by writer. Each crop is a derivative of a permissively-licensed upstream scan in HeOCR/hash, with per-image rights inherited and attribution recorded. The index is line-oriented JSON (JSONL). Release 0.0.0-rc contains 48 per-letter image entries drawn from 2 verified writers (18 LicenseRef-Public-Domain-Israel, 30 PDM-1.0).
 authors:
 - name: Shay Palachy-Affek
 version: 0.0.0-rc

diff --git a/LICENSE.md b/LICENSE.md
@@ -1,9 +1,9 @@
 # Licensing Policy
 
 This repository is structured for compound licensing — the same model used
-by [HeOCR/public-domain-hand-written-hebrew-scans][upstream].
+by [HeOCR/hash][upstream].
 
-[upstream]: https://github.com/HeOCR/public-domain-hand-written-hebrew-scans
+[upstream]: https://github.com/HeOCR/hash
 
 ## Repository-authored metadata
 
@@ -32,7 +32,7 @@ material is separately released under compatible terms.
 ## Per-letter image crops
 
 Per-letter image bytes are **derivatives** of upstream scans hosted in
-[HeOCR/public-domain-hand-written-hebrew-scans][upstream]. They are not
+[HeOCR/hash][upstream]. They are not
 automatically covered by the metadata license. Each crop carries its own
 entry-level rights record in `data/index/entries.jsonl`:
 

diff --git a/NOTICE.md b/NOTICE.md
@@ -4,7 +4,7 @@ This file is generated by `scripts/generate_release_artifacts.py` from `data/ind
 
 Repository-authored metadata is dedicated to the public domain under CC0 1.0 Universal. See [`LICENSE`](LICENSE) and [`LICENSE.md`](LICENSE.md) for the full compound-licensing policy.
 
-Per-letter image crops are derivatives of upstream scans in [HeOCR/public-domain-hand-written-hebrew-scans](https://github.com/HeOCR/public-domain-hand-written-hebrew-scans) and carry per-entry rights inherited from the source page. The entries listed below carry a license that requires attribution (currently CC-BY-4.0, CC-BY-SA-4.0). Anyone redistributing or reusing these crops must keep the listed credit and link to the source page on which the rights claim was verified.
+Per-letter image crops are derivatives of upstream scans in [HeOCR/hash](https://github.com/HeOCR/hash) and carry per-entry rights inherited from the source page. The entries listed below carry a license that requires attribution (currently CC-BY-4.0, CC-BY-SA-4.0). Anyone redistributing or reusing these crops must keep the listed credit and link to the source page on which the rights claim was verified.
 
 - Corpus release: `0.0.0-rc`
 - Released at (corpus state): `2026-05-13T21:37:46Z`

diff --git a/README.md b/README.md
@@ -7,7 +7,7 @@ cut from different scans by that writer.
 
 This repository is the downstream of:
 
-- [HeOCR/public-domain-hand-written-hebrew-scans][upstream] — the
+- [HeOCR/hash][upstream] (HASH — Hebrew Archive of Scanned Handwriting) — the
   canonical, permissively-licensed source of page-level scans. Every
   entry here cites its upstream scan.
 - [HeOCR/hletterscriptgen][gen] — the framework that turns page scans
@@ -19,7 +19,7 @@ The intended downstream consumers are synthetic-document generators
 corpora they feed into ([HeOCR/HeOCRsynth][heocrsynth],
 [HeOCR/HeOCR][heocr]).
 
-[upstream]: https://github.com/HeOCR/public-domain-hand-written-hebrew-scans
+[upstream]: https://github.com/HeOCR/hash
 [gen]: https://github.com/HeOCR/hletterscriptgen
 [syngen]: https://github.com/HeOCR/hocrsyngen
 [heocrsynth]: https://github.com/HeOCR/HeOCRsynth

diff --git a/datapackage.json b/datapackage.json
@@ -5,7 +5,7 @@
       "title": "Shay Palachy-Affek"
     }
   ],
-  "description": "Per-letter image crops of handwritten Hebrew letters, grouped into sets by writer. Each crop is a derivative of a permissively-licensed upstream scan in HeOCR/public-domain-hand-written-hebrew-scans, with per-image rights inherited and attribution recorded. The index is line-oriented JSON (JSONL).",
+  "description": "Per-letter image crops of handwritten Hebrew letters, grouped into sets by writer. Each crop is a derivative of a permissively-licensed upstream scan in HeOCR/hash, with per-image rights inherited and attribution recorded. The index is line-oriented JSON (JSONL).",
   "homepage": "https://github.com/HeOCR/hletterscript",
   "keywords": [
     "Hebrew",
@@ -113,7 +113,7 @@
     }
   },
   "title": "Hebrew Handwritten Per-Letter Image Dataset",
-  "upstream_repo": "https://github.com/HeOCR/public-domain-hand-written-hebrew-scans",
+  "upstream_repo": "https://github.com/HeOCR/hash",
   "version": "0.0.0-rc",
   "version_released_date": "2026-05-12"
 }
diff --git a/docs/dataset_structure.md b/docs/dataset_structure.md
@@ -10,7 +10,7 @@ document or scan written by that writer.
 
 The corpus is the *downstream* product of two upstream things:
 
-- [HeOCR/public-domain-hand-written-hebrew-scans] is the canonical source
+- [HeOCR/hash] is the canonical source
   of page-level scans. Every per-letter image entry in this repo cites
   the upstream scan (`source_id`, `entry_id`, `sha256`) it was cut from.
 - [HeOCR/hletterscriptgen] is the framework that turns those page scans
@@ -21,7 +21,7 @@ The intended downstream consumers are synthetic-document generators
 ([HeOCR/hocrsyngen]) and the synthetic / real Hebrew handwriting datasets
 they feed into ([HeOCR/HeOCRsynth], [HeOCR/HeOCR]).
 
-[HeOCR/public-domain-hand-written-hebrew-scans]: https://github.com/HeOCR/public-domain-hand-written-hebrew-scans
+[HeOCR/hash]: https://github.com/HeOCR/hash
 [HeOCR/hletterscriptgen]: https://github.com/HeOCR/hletterscriptgen
 [HeOCR/hocrsyngen]: https://github.com/HeOCR/hocrsyngen
 [HeOCR/HeOCRsynth]: https://github.com/HeOCR/HeOCRsynth
@@ -155,7 +155,7 @@ operational form of this rule.
 
 Every per-letter image is a **crop / derivative** of an upstream scan whose
 rights have already been recorded in
-`public-domain-hand-written-hebrew-scans/data/index/entries.jsonl`.
+`hash/data/index/entries.jsonl`.
 Repository policy:
 
 - **Repository-authored metadata** in this repo is dedicated to the public

diff --git a/schemas/entry.schema.json b/schemas/entry.schema.json
@@ -2,7 +2,7 @@
   "$schema": "https://json-schema.org/draft/2020-12/schema",
   "$id": "https://github.com/HeOCR/hletterscript/schemas/entry.schema.json",
   "title": "Handwritten Hebrew Per-Letter Image Entry",
-  "description": "One row per cropped per-letter image. Each entry is a derivative of a specific upstream scan in HeOCR/public-domain-hand-written-hebrew-scans.",
+  "description": "One row per cropped per-letter image. Each entry is a derivative of a specific upstream scan in HeOCR/hash.",
   "type": "object",
   "required": [
     "entry_id",
@@ -175,7 +175,7 @@
     },
     "upstream": {
       "type": "object",
-      "description": "Reference back to the scan in HeOCR/public-domain-hand-written-hebrew-scans this crop was extracted from. The upstream repository URL is recorded once in scripts/release_recipe.json (`upstream_repo`); it is not duplicated on every entry.",
+      "description": "Reference back to the scan in HeOCR/hash this crop was extracted from. The upstream repository URL is recorded once in scripts/release_recipe.json (`upstream_repo`); it is not duplicated on every entry.",
       "required": ["source_id", "entry_id", "sha256", "commit", "bbox"],
       "additionalProperties": false,
       "properties": {

diff --git a/scripts/generate_release_artifacts.py b/scripts/generate_release_artifacts.py
@@ -234,7 +234,7 @@ def _notice_stanza(
 for the full compound-licensing policy.
 
 Per-letter image crops are derivatives of upstream scans in \
-[HeOCR/public-domain-hand-written-hebrew-scans]({upstream_repo_url}) and \
+[HeOCR/hash]({upstream_repo_url}) and \
 carry per-entry rights inherited from the source page. The entries \
 listed below carry a license that requires attribution (currently \
 {license_set}). Anyone redistributing or reusing these crops must keep \

diff --git a/scripts/release_recipe.json b/scripts/release_recipe.json
@@ -4,10 +4,10 @@
   "version": "0.0.0-rc",
   "version_released_date": "2026-05-12",
   "initial_release_date": "2026-05-12T00:00:00Z",
-  "description": "Per-letter image crops of handwritten Hebrew letters, grouped into sets by writer. Each crop is a derivative of a permissively-licensed upstream scan in HeOCR/public-domain-hand-written-hebrew-scans, with per-image rights inherited and attribution recorded. The index is line-oriented JSON (JSONL).",
+  "description": "Per-letter image crops of handwritten Hebrew letters, grouped into sets by writer. Each crop is a derivative of a permissively-licensed upstream scan in HeOCR/hash, with per-image rights inherited and attribution recorded. The index is line-oriented JSON (JSONL).",
   "homepage": "https://github.com/HeOCR/hletterscript",
   "repository_code": "https://github.com/HeOCR/hletterscript",
-  "upstream_repo": "https://github.com/HeOCR/public-domain-hand-written-hebrew-scans",
+  "upstream_repo": "https://github.com/HeOCR/hash",
   "authors": [
     {"name": "Shay Palachy-Affek", "role": "maintainer"}
   ],

diff --git a/scripts/review_crops.py b/scripts/review_crops.py
@@ -113,7 +113,7 @@ def _build_html(entries: list[dict], upstream_root: Path | None) -> str:
 <div class="scan-block">
   <h2>Upstream scan: <code>{upstream_entry_id}</code></h2>
   <p class="warn">Upstream scan not found locally. Run with
-  <code>--upstream-path /path/to/public-domain-hand-written-hebrew-scans</code>
+  <code>--upstream-path /path/to/hash</code>
   to display it.</p>
 </div>
 """
@@ -391,7 +391,7 @@ def do_POST(self):
 def main() -> None:
     ap = argparse.ArgumentParser(description="Serve a crop-review page locally.")
     ap.add_argument("--upstream-path", metavar="PATH",
-                    help="Path to a clone of HeOCR/public-domain-hand-written-hebrew-scans")
+                    help="Path to a clone of HeOCR/hash")
     ap.add_argument("--output", metavar="FILE",
                     help="Write the HTML to this file instead of serving it")
     ap.add_argument("--port", type=int, default=8765,

diff --git a/scripts/validate_indexes.py b/scripts/validate_indexes.py
@@ -392,7 +392,7 @@ def _load_upstream_entries(upstream_root: Path) -> dict[str, dict[str, Any]]:
         raise SystemExit(_err(
             upstream_entries_path, None, None,
             "upstream entries.jsonl not found; --upstream-path must point at a clone of "
-            "public-domain-hand-written-hebrew-scans",
+            "hash",
         ))
     by_id: dict[str, dict[str, Any]] = {}
     with upstream_entries_path.open("r", encoding="utf-8") as handle:
@@ -493,7 +493,7 @@ def main() -> None:
         type=Path,
         default=None,
         help=(
-            "Path to a local clone of HeOCR/public-domain-hand-written-hebrew-scans. "
+            "Path to a local clone of HeOCR/hash. "
             "When set, the validator additionally cross-checks each entry's "
             "upstream.sha256 against the upstream file record and verifies "
             "upstream.bbox fits inside the upstream scan dimensions."