awaforge

oconvPptxToIr

AGPL-3.0-onlyRelease 2L3

.pptx (@awacloud/ooxml's pptx.read() result) → oconv-ir/v1, tier 1 — one heading + N paragraphs per slide.

Module oconvPptxToIr (oconvPptxToIr) | Source packages/front/office/oconv/src/read/pptx-to-ir.js | Deps oconvIr, pptx | Worker-safe yes

Pure transformation over an already-parsed pptx.read(bytes) structure. There is no "section" node in the frozen IR vocabulary, so each slide contributes a flat run of top-level blocks in slide order.

Resolve

import { ModuleRuntime } from '@awacloud/fw/core/runtime.js';
import { fw_require, modules } from '@awacloud/oconv';

const runtime = new ModuleRuntime();
runtime.registerAll(fw_require);
runtime.registerAll(modules);
const pptx = runtime.resolve('pptx');
const { pptxToIr } = runtime.resolve('oconvPptxToIr');

API

Member Signature Returns Throws
pptxToIr (readResult: object, opts?: {includeNotes?: boolean}) => {ir, losses} ir — an oconv-ir/v1 document; losses — {code, detail}[] —

Examples

Convert a .pptx read result to IR

const { ir, losses } = pptxToIr(pptx.read(bytes), { includeNotes: false });
ir.kind;   // 'document'

Notes

  • Title — the first shape whose placeholder.type is 'title' or 'ctrTitle'; a title shape present but carrying no non-empty text counts as "no title" (loss slides/untitled).
  • Body — every other shape carrying a txBody, in shape order, one IR paragraph per non-empty text paragraph.
  • includeNotes is a documented no-op for pptx: pptx.read()'s public return shape never surfaces speaker notes (no notes field on a slide, no relationship resolution to notesSlide<n>.xml), so this reader never emits a notes blockquote and never records slides/notes-omitted for pptx — the option is accepted only for frozen shape parity with oconvOdpToIr, whose underlying format does expose notes.
  • Loss codes this module emits: slides/untitled, slides/media-dropped (a picture/table/chart/graphicFrame shape, detail = its type), slides/notes-omitted (never actually reachable for pptx — see above).
  • Layout and animations (layoutRef, transitions) are permanently out of scope and never reach a per-node loss — pptx.read()'s public result does not even model them.

See also