Re-extract: full statement+proof, normalized schema, provenance

README.md

@@ -18,7 +18,6 @@ size_categories:
   - 10K<n<100K
 dataset_info:
   features:
-  - {name: fact, dtype: string}
   - {name: statement, dtype: string}
   - {name: proof, dtype: string}
   - {name: type, dtype: string}
@@ -58,16 +57,15 @@ Structured dataset from the Mathematical Components library (MathComp) for Coq.
 
 | Column | Type | Description |
 |--------|------|-------------|
-| fact | string | Verbatim declaration with the leading keyword removed: signature and body/proof joined |
-| statement | string | Signature with the leading keyword removed (verbatim slice) |
-| proof | string | Verbatim proof/body, empty if none |
+| statement | string | Declaration signature/claim with the leading keyword removed (verbatim slice); the full declaration minus its proof |
+| proof | string | Verbatim proof/body, empty if the declaration has none |
 | type | string | Declaration keyword |
 | symbolic_name | string | Declaration identifier |
 | library | string | Sub-library |
 | filename | string | Repository-relative source path |
 | imports | list[string] | File-level `Require`/`Import` modules |
 | deps | list[string] | Intra-corpus identifiers referenced |
-| docstring | string | Preceding documentation comment, null if absent |
+| docstring | string | Preceding documentation comment, empty if absent |
 | line_start | int | First source line |
 | line_end | int | Last source line |
 | has_proof | bool | Whether a proof block was captured |
@@ -113,7 +111,7 @@ Structured dataset from the Mathematical Components library (MathComp) for Coq.
 
 ## Example
 
-```coq
+```
 floor_subproof n :
   if n \is real_num then n%:~R <= n < (n + 1)%:~R else n == 0.
 Proof. by rewrite num_real !intz ltzD1 lexx. Qed.
@@ -123,8 +121,9 @@ Proof. by rewrite num_real !intz ltzD1 lexx. Qed.
 
 ## Use
 
-Statement and proof are available both joined (`fact`) and split (`statement`, `proof`) for
-proof-term modeling, autoformalization, retrieval, and dependency analysis via `deps`.
+Each declaration is split into a `statement` (signature/claim) and a `proof` (body) that are disjoint
+and together form the complete declaration, for proof modeling, autoformalization, retrieval, and
+dependency analysis via `deps`.
 
 ## Citation
 

data/train-00000-of-00001.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:13b2df14fdbb263f6fa9ec46ccfe3d0a93bc042afdf139569e59306704c4ddac
-size 4602650
+oid sha256:7b6e1167f1377d07bd20f5009c12fbbff02b8d18cecc4a08881b7de90f35a669
+size 2710162