PDF Reader And Objects

muhammara.createReader(input, options) returns a PDFReader for a file path or compatible random-access read stream.

Pass { password: "..." } as the options argument to open an encrypted PDF. Either its user or owner password is accepted. Call isEncrypted() after creation when the input may be encrypted:

var reader = muhammara.createReader("encrypted.pdf", {
  password: "open-password",
});

The reader provides document information with getPDFLevel, getPagesCount, getTrailer, and isEncrypted; page access with parsePage, parsePageDictionary, and getPageObjectID; and object access with parseNewObject, queryDictionaryObject, and queryArrayObject.

extractPageText(pageIndex, limits?) returns content-stream text operations in drawing order. Each PDFTextElement includes text decoded to Unicode through the active font, the raw character codes as content, fontResource, fontSize, and a six-value textMatrix; it does not compute glyph bounds. See Find Text Positions for how text is decoded.

extractPageContentItems(pageIndex, limits?) returns every direct content-stream operation that puts a mark on the page as { type, operation }, which answers "is this page blank?" without extracting text.

Both accept the same optional limits object. Omitted fields keep the built-in default, and values above it are clamped down, so a caller can tighten the extraction budget but never raise it past the ceiling the extractor enforces. See Find Text Positions for the field table.

Every method that takes a page index or object ID — parsePage, parsePageDictionary, getPageObjectID, extractPageText, extractPageContentItems, parseNewObject, and getXrefEntry — requires a non-negative integer below 2^32 and throws a TypeError otherwise. Negative and fractional values are rejected rather than coerced, so parsePage(-1) throws instead of reading a wildly out-of-range page.

Parsed PDFObject values expose getType, conversion methods such as toPDFDictionary() and toPDFArray(), and scalar conversion through toNumber() and toString(). A conversion to a different object type returns undefined, so check the result or getType() first. A dictionary provides exists, queryObject, and toJSObject; an array provides getLength, queryObject, and toJSArray. Use reader query helpers when an entry may be an indirect reference and must be resolved.

end() releases the reader and the file handle behind it. Call it once every parsed object and stream has been consumed; until then the input file stays open, and Windows refuses to rename or delete it.

var reader = muhammara.createReader("input.pdf");
var pageCount = reader.getPagesCount();
var firstPage = reader.parsePage(0);
reader.end();

For a complete reader workflow, see Read PDFs.

getXrefSize, getXrefPosition, getParserStream, startReadingFromStream, startReadingFromStreamForPlainCopying, startReadingObjectsFromStream, and startReadingObjectsFromStreams expose raw parser internals for specialized low-level processing. They are not stable general-purpose reading workflows; use the page and object methods above unless you need to build a PDF parser integration. The byte readers they return deliver each read(amount) as a Buffer.