Arabic root frequency across the library
shamela_root_statsانتشار جذر في المكتبة
Profile how widely an Arabic root spreads across the DOWNLOADED library, aggregated by category / Hijri century / book / author. Runs ONE morphological (AlKhalil) page search for the root — all derived forms are counted (صابر/يصبر/اصطبار for صبر) — and returns the DISTRIBUTION only, not snippets. total_hits is EXACT; the by-category/century/book/author breakdown is built from at most 5,000 top-scoring hits (COVERAGE_CAP), so when coverage_capped is true the bucket counts are floors and shares are indicative. Morphology accuracy on classical Arabic is ~0.80 — read counts as reach, not exact tallies. Pass scope to profile a slice. Examples: shamela_root_stats({root:'صبر'}), shamela_root_stats({root:'رحم', scope:{category_ids:[17]}}).
Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| mode | No | 'literal' counts pages carrying the WORD as written (plus the ordinary spelling folds). 'morphological' expands it to every AlKhalil root reading and counts those — which is far broader than most callers expect: «الاستصناع» reports 393,498 pages across 7,405 books morphologically against 1,510 literally, an answer that says every shelf holds the topic. Use 'literal' to size a topic before planning retrieval; use 'morphological' to trace a ROOT through its derivations. | morphological |
| root | Yes | Arabic root or base word to profile (e.g. «صبر»، «رحم»، «علم»). Under mode='morphological' it is matched via the AlKhalil analyzer, so all derived forms (صابر، يصبر، اصطبار…) are counted, not just the literal string. | |
| scope | No | Restrict the profile to specific books, authors, categories, or a Hijri date range. Use shamela_list_categories / shamela_resolve to find IDs. Omit to profile the whole downloaded library. | |
| response_format | No | Output format for the human-readable text. 'markdown' (default) is best for direct display; 'json' returns the same data as a single JSON code block; 'none' skips the rendered text entirely and returns a one-line pointer, leaving `structuredContent` as the whole answer. The structured data is returned in `structuredContent` regardless of this setting, so a caller that parses it should pass 'none' — the character budget is measured over BOTH channels, and the unread copy is otherwise spending most of it. Requires the 'response_format.none' capability (see shamela_brief). | markdown |