P33 the rijal registries v1 (five seed texts)

The law on this layer's face:
PX1-F3: the paragraph-proxy ERROR DIRECTION INVERTS between works (panel-measured: Thiqat ~17x UNDER, Majruhin ~2x OVER, Tusi-Rijal under) - the field is renamed text_blocks_parsed and NO entry-count claim stands until the per-work roll parsers run; Tusi-Rijal (its rolls are line-coded (ain)-entries - the roll parser is the refinement, declared); segmentation classes DECLARED per work (Ikhtiyar by its own numbers; others by name-paragraph PROXY - the finer parser is the bench's refinement); the jarh/ta'dil vocabulary censused on verbatim text; name-strings never merged into persons (I-1: the identification is scholars' work); the Ghada'iri 226 stand in the ocean, genre-tagged, joined by reference
What this means for a reader: every number below was counted by a program from the sealed sources and can be re-derived by anyone running the same engine; nothing here judges a scholar or a translator - it makes the text's own facts checkable.
MeasureValue
built 2026-09-04T11:58:34 · source_sha fbc8744a98b36edb · AL-WAQF / P31 AL-MIZAN · machine-scholar layer, human review invited
← AL-WAQF