Method format

src/org/antlr/codebuff/Formatter.java:123–167

Format the document. Does not affect/alter doc.

(InputDocument doc, boolean collectAnalysis)

	/** Format the document. Does not affect/alter doc. */
	public String format(InputDocument doc, boolean collectAnalysis) throws Exception {
		if ( testDoc!=null ) throw new IllegalArgumentException("can't call format > once");
		// for debugging we need a map from original token with actual line:col to tree node. used by token analysis
		originalDoc = doc;
		originalTokenToNodeMap = indexTree(doc.tree);
		originalTokens = doc.tokens;

		this.testDoc = InputDocument.dup(doc); // make copy of doc, getting new tokens, tree
		output = new StringBuilder();
		this.realTokens = getRealTokens(testDoc.tokens);
		// squeeze out ws and kill any line/col info so we can't use ground truth by mistake
		wipeCharPositionInfoAndWhitespaceTokens(testDoc.tokens); // all except for first token
		wsClassifier = new kNNClassifier(corpus, wsFeatures, corpus.injectWhitespace);
		hposClassifier = new kNNClassifier(corpus, hposFeatures, corpus.hpos);

		analysis = new Vector<>(testDoc.tokens.size());
		analysis.setSize(testDoc.tokens.size());

		// make an index on the duplicated doc tree with tokens missing line:col info
		if ( tokenToNodeMap == null ) {
			tokenToNodeMap = indexTree(testDoc.tree);
		}

		WritableToken firstToken = (WritableToken)testDoc.tokens.getNextRealToken(-1);

		String prefix = originalTokens.getText(Interval.of(0, firstToken.getTokenIndex())); // gets any comments in front + first real token
		charPosInLine = firstToken.getCharPositionInLine()+firstToken.getText().length()+1; // start where first token left off
		line = Tool.count(prefix, '\n') + 1;
		output.append(prefix);

		// first identify oversize lists with separators
		IdentifyOversizeLists splitter = new IdentifyOversizeLists(corpus, testDoc.tokens, tokenToNodeMap);
		ParseTreeWalker.DEFAULT.walk(splitter, testDoc.tree);
		tokenToListInfo = splitter.tokenToListInfo;

		realTokens = getRealTokens(testDoc.tokens);
		for (int i = Trainer.ANALYSIS_START_TOKEN_INDEX; i<realTokens.size(); i++) { // can't process first token
			int tokenIndexInStream = realTokens.get(i).getTokenIndex();
			processToken(i, tokenIndexInStream, collectAnalysis);
		}

		releaseMemory();

		return output.toString();
	}

	public float getWSEditDistance() throws Exception {
		List<Token> wsTokens = filter(originalTokens.getTokens(),