Normalize and restore PROBE Gene2Vec package

Restore the complete probe_gene2vec_uniprot package. Reduce genes.tsv to its ordered Ensembl gene ID column, compact metadata.json, update README provenance, and restore embeddings.npz byte-identically.

data/embeddings/probe_gene2vec_uniprot/README.md

@@ -0,0 +1,177 @@
+# Embedding name: PROBE-GENE2VEC-UNIPROT
+
+- Modality: Protein-associated representation
+- Algorithm: Official PROBE human UniProt representation table, standardized by UniProt-to-master mapping
+- Embedding source URL: https://huggingface.co/spaces/HUBioDataLab/PROBE
+- Original model/code URL: https://github.com/HUBioDataLab/PROBE
+
+## Public files
+
+- `embeddings.npz`: compressed NumPy archive containing the embedding matrix under key `embeddings`.
+- `genes.tsv`: row metadata table. Row order matches the embedding matrix.
+- `metadata.json`: machine-readable provenance and processing metadata.
+
+## Matrix
+
+- Shape: `17,564 genes × 200 dimensions`
+- Dtype: `float32`
+
+
+
+## Detailed processing notes
+
+### Source and interpretation
+
+This embedding is one PROBE human UniProt representation table standardized into the shared gene-embedding resource.
+
+This is a Gene2Vec-associated UniProt-keyed representation from the PROBE archive.
+
+The source file used for this component was `GENE2VEC_UNIPROT_HUMAN.csv` from the official PROBE human UniProt representation archive. The raw vectors were not regenerated locally. The local work here was identifier cleaning, UniProt-to-gene mapping, duplicate handling, QC, and conversion into the shared public resource format.
+
+### Provenance fields
+
+    source_name: PROBE-GENE2VEC-UNIPROT
+    original_model_or_method: PROBE human UniProt representation archive
+    input_data_type: official PROBE human UniProt CSV representation table
+    input_source_file: GENE2VEC_UNIPROT_HUMAN.csv
+    source_download_or_app_page: https://huggingface.co/spaces/HUBioDataLab/PROBE
+    source_code_page: https://github.com/HUBioDataLab/PROBE
+    raw_source_folder: ~/data/raw_embeddings/probe_official/human_representation_vectors/revision-1/
+    processing_script: ~/scripts/process_embeddings/process_probe_all_v1_1.py
+    preferred_batch_summary_report: ~/reports/mapping_reports/probe_all_v1_1_summary.tsv
+    source_identifier_type: UniProt Entry accession
+    mapping_identifier_type: UniProt accession matched to enriched master_gene_table.csv UniProt fields
+    source_rows: 18,407
+    final_mapped_genes: 17,564
+    model_or_feature_dimension: 200
+    training_or_finetuning: none
+    final_embedding_dtype: float32
+
+### Mapping and processing decisions
+
+Each PROBE CSV uses a UniProt accession in the `Entry` column. Source entries were cleaned to their canonical accession base by removing isoform suffixes when present.
+
+Cleaned UniProt accessions were mapped to master_gene_table.csv using `uniprot_id`, `hgnc_uniprot_ids`, and `map_uniprot_ids_all`. Only unique UniProt-to-master mappings were accepted. Ambiguous UniProt accessions were excluded. Duplicate source proteins mapping to the same final master-table gene were averaged within that gene.
+
+The exact underlying UniProt/protein sequence database snapshot is inherited from the PROBE archive and should be checked against the official PROBE release documentation if a strict sequence-database timestamp is required.
+
+### Processing workflow
+
+For this PROBE embedding:
+
+    1. Read `GENE2VEC_UNIPROT_HUMAN.csv` from the PROBE human representation archive.
+    2. Use the `Entry` column as the source UniProt accession.
+    3. Remove isoform suffixes where present to obtain the canonical UniProt accession base.
+    4. Map source UniProt accessions to master_gene_table.csv through enriched UniProt fields.
+    5. Exclude unmapped and ambiguous UniProt accessions.
+    6. Average duplicate source rows only when they map to the same final master-table gene.
+    7. Save the final matrix as float32.
+
+### Important implementation decisions
+
+The PROBE vectors were taken directly from the official archive rather than recomputed. This keeps the PROBE batch aligned with the released human UniProt representation files.
+
+All PROBE embeddings use the same mapping logic, so differences between PROBE outputs reflect the source representation type rather than different gene-mapping rules.
+
+### Summary of final mapping:
+
+    Source identifiers / source rows: 18,407
+    Dimensions: 200
+    Final mapped genes: 17,564
+    Master table rows: 20,097
+    Master coverage: 87.40%
+    Mapping strategy: UniProt Entry accession mapping through enriched master table
+    Ambiguous UniProt accessions: excluded
+    Duplicate final mappings averaged: yes
+
+<!-- BEGIN CATALOG NORMALIZATION: LEGACY METADATA -->
+
+## Processing and provenance metadata
+
+This section preserves information moved out of `metadata.json` during catalog normalization. The loading-facing metadata is intentionally compact; detailed source, mapping, processing, quality-control, and reproducibility records remain here.
+
+Source catalog commit: `b012c9b6cb302f5b559a8ab3dba3aabf9f6b3e91`.
+
+### Canonicalized field names
+
+| Previous field | Compact metadata field |
+| --- | --- |
+| `embedding_name` | `display_name` |
+| `source_identifier_type` | `original_identifier_type` |
+| `mapping_strategy` | `mapping_method` |
+| `num_genes` | `n_genes` |
+| `embedding_dimension` | `n_dimensions` |
+| `matrix_shape[0]` | `n_genes` |
+| `matrix_shape[1]` | `n_dimensions` |
+| `coverage.master_gene_coverage_fraction` | `master_coverage_fraction` |
+| `original_model_or_method` | `original_method` |
+| `download_source` | `input_data_source` |
+
+### Derived and verified fields
+
+- `species` was standardized to `Homo sapiens` from the validated human Ensembl `ENSG` identifier namespace.
+- `dtype` and `npz_key` were read directly from the validated `embeddings.npz` archive.
+- `publication` uses the verified free-access publication link `https://rdcu.be/cJAKN` for *Learning functional properties of proteins with language models*.
+
+### Legacy processing and provenance record
+
+```json
+{
+  "all_zero_rows": {
+    "count": 0,
+    "documented": true,
+    "expected": false,
+    "reason": "No exact all-zero rows were detected in embeddings.npz.",
+    "report_file": null
+  },
+  "counts": {
+    "ambiguous_source_rows": 169,
+    "duplicate_final_mappings_averaged": 27,
+    "final_dimensions": 200,
+    "final_mapped_genes": 17564,
+    "mapped_source_rows_before_duplicate_collapse": 17593,
+    "master_rows": 20097,
+    "source_dimensions": 200,
+    "source_rows": 18407,
+    "unmapped_source_rows": 645
+  },
+  "coverage": {
+    "master_gene_coverage_fraction": 0.8739612877543912,
+    "source_row_mapping_fraction": 0.9557776932688651
+  },
+  "download_source": "Official PROBE human_representation_vectors.tar.gz archive from Zenodo record 5795850 plus official HUBioDataLab/PROBE repository",
+  "embedding_dimension": 200,
+  "embedding_generated_by": "PROBE authors / original PROBE human representation archive",
+  "embedding_id": "probe_gene2vec_uniprot",
+  "embedding_name": "probe_gene2vec_uniprot_v1_1",
+  "former_processing_embedding_id": "probe_gene2vec_uniprot_v1_1",
+  "mapping_method_counts": {
+    "uniprot_accession_exact_unique": 17564
+  },
+  "mapping_strategy": "Exact UniProt accession mapping to master_gene_table_v1_1_enriched.csv using uniprot_id, hgnc_uniprot_ids, and map_uniprot_ids_all. Only unique mappings are accepted. Duplicate source rows mapping to the same final master gene are averaged.",
+  "matrix_shape": [
+    17564,
+    200
+  ],
+  "modality": "protein sequence / protein representation from PROBE human UniProt archive",
+  "num_genes": 17564,
+  "original_model_or_method": "GENE2VEC",
+  "public_file_structure": {
+    "embeddings": "embeddings.npz",
+    "genes": "genes.tsv",
+    "metadata": "metadata.json",
+    "readme": "README.md"
+  },
+  "sequence_and_reference_provenance": {
+    "input_sequence_source": "PROBE human UniProt representation archive",
+    "input_sequence_type": "protein sequence-derived representation",
+    "sequence_reference_status": "PROBE archive provenance recorded; exact UniProt release to verify from original PROBE documentation if needed",
+    "source_identifier": "UniProt accession in Entry column",
+    "species": "human"
+  },
+  "source_embedding": "PROBE_GENE2VEC",
+  "source_identifier_type": "UniProt accession"
+}
+```
+
+<!-- END CATALOG NORMALIZATION: LEGACY METADATA -->

data/embeddings/probe_gene2vec_uniprot/embeddings.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a52a9f963c9fe041ee0b89f305d0f7bdb9c267b2db48702578bf2d20ed7fafb8
+size 13029054

data/embeddings/probe_gene2vec_uniprot/metadata.json

@@ -1,56 +1,17 @@
 {
-  "embedding_name": "probe_gene2vec_uniprot_v1_1",
-  "source_embedding": "PROBE_GENE2VEC",
+  "embedding_id": "probe_gene2vec_uniprot",
+  "display_name": "probe_gene2vec_uniprot_v1_1",
   "modality": "protein sequence / protein representation from PROBE human UniProt archive",
-  "source_identifier_type": "UniProt accession",
-  "download_source": "Official PROBE human_representation_vectors.tar.gz archive from Zenodo record 5795850 plus official HUBioDataLab/PROBE repository",
-  "original_model_or_method": "GENE2VEC",
+  "species": "Homo sapiens",
+  "original_identifier_type": "UniProt accession",
+  "mapping_method": "Exact UniProt accession mapping to master_gene_table_v1_1_enriched.csv using uniprot_id, hgnc_uniprot_ids, and map_uniprot_ids_all. Only unique mappings are accepted. Duplicate source rows mapping to the same final master gene are averaged.",
+  "n_genes": 17564,
+  "n_dimensions": 200,
+  "dtype": "float32",
+  "npz_key": "embeddings",
+  "master_coverage_fraction": 0.8739612877543912,
+  "original_method": "GENE2VEC",
+  "publication": "https://rdcu.be/cJAKN",
   "embedding_generated_by": "PROBE authors / original PROBE human representation archive",
-  "sequence_and_reference_provenance": {
-    "input_sequence_type": "protein sequence-derived representation",
-    "input_sequence_source": "PROBE human UniProt representation archive",
-    "source_identifier": "UniProt accession in Entry column",
-    "species": "human",
-    "sequence_reference_status": "PROBE archive provenance recorded; exact UniProt release to verify from original PROBE documentation if needed"
-  },
-  "mapping_strategy": "Exact UniProt accession mapping to master_gene_table_v1_1_enriched.csv using uniprot_id, hgnc_uniprot_ids, and map_uniprot_ids_all. Only unique mappings are accepted. Duplicate source rows mapping to the same final master gene are averaged.",
-  "counts": {
-    "source_rows": 18407,
-    "source_dimensions": 200,
-    "mapped_source_rows_before_duplicate_collapse": 17593,
-    "unmapped_source_rows": 645,
-    "ambiguous_source_rows": 169,
-    "final_mapped_genes": 17564,
-    "final_dimensions": 200,
-    "master_rows": 20097,
-    "duplicate_final_mappings_averaged": 27
-  },
-  "coverage": {
-    "source_row_mapping_fraction": 0.9557776932688651,
-    "master_gene_coverage_fraction": 0.8739612877543912
-  },
-  "mapping_method_counts": {
-    "uniprot_accession_exact_unique": 17564
-  },
-  "embedding_id": "probe_gene2vec_uniprot",
-  "former_processing_embedding_id": "probe_gene2vec_uniprot_v1_1",
-  "public_file_structure": {
-    "embeddings": "embeddings.npz",
-    "genes": "genes.tsv",
-    "metadata": "metadata.json",
-    "readme": "README.md"
-  },
-  "num_genes": 17564,
-  "embedding_dimension": 200,
-  "matrix_shape": [
-    17564,
-    200
-  ],
-  "all_zero_rows": {
-    "count": 0,
-    "documented": true,
-    "expected": false,
-    "reason": "No exact all-zero rows were detected in embeddings.npz.",
-    "report_file": null
-  }
+  "input_data_source": "Official PROBE human_representation_vectors.tar.gz archive from Zenodo record 5795850 plus official HUBioDataLab/PROBE repository"
 }