FileProcessor.java

package com.cadrlife.devsearch.agent.service.analysis;

import com.google.common.collect.ImmutableMap;
import org.elasticsearch.client.Client;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

import javax.inject.Inject;
import java.nio.file.Path;
import java.util.*;
import java.util.regex.Matcher;
import java.util.regex.Pattern;

import static org.apache.commons.lang.StringUtils.remove;

public class FileProcessor {
    JavaReferenceFinder javaReferenceFinder;
    @Inject
    public FileProcessor(JavaReferenceFinder javaReferenceFinder) {
        this.javaReferenceFinder = javaReferenceFinder;
    }
    private static final Logger LOG = LoggerFactory.getLogger(FileProcessor.class);

    public void addSourceAnalysisFields(ImmutableMap.Builder<String,Object> mapBuilder, Path filePath, String content) {
        String fileName = filePath.getFileName().toString();
        if (fileName.endsWith(".java") || fileName.endsWith(".groovy")) {
            LOG.debug("Analyzing file {}", filePath);
            mapBuilder.putAll(computeJavaStuff(content, filePath).asMap());
        }
    }

    JavaAnalysis computeJavaStuff(String content, Path filePath) {
        JavaAnalysis analysis = new JavaAnalysis();
        Set<String> imports = new TreeSet<String>();

        Scanner scanner = new Scanner(content);
        int lineNumber = 0;
        while( scanner.hasNextLine() ) {
            lineNumber++;
            String trimmedLine = scanner.nextLine().trim();
            if (trimmedLine.startsWith("package ")) {
                String filePackage = remove(remove(trimmedLine, "package "), ";").trim();
                String baseName = com.google.common.io.Files.getNameWithoutExtension(filePath.getFileName().toString());
                LOG.debug("Detected package {}", filePackage);
                LOG.debug("Detected javaName {}", baseName);
                analysis.setJavaPackage(filePackage);
                analysis.setName(baseName);
            }
            if (trimmedLine.startsWith("import ")) {
                String foundImport = remove(remove(remove(remove(trimmedLine, "import "), ";"), "static "), ".*").trim();
                LOG.debug("Detected import {}", foundImport);
                analysis.addImport(lineNumber, foundImport);
            } else {
                Matcher m  = Pattern.compile("\\b(([a-z0-9]+[\\.])+[A-Z]+\\w+)").matcher(trimmedLine);
                if( m.find()) {
                    String foundImport = m.group(0);
                    LOG.debug("Detected import with weird regex {}", foundImport);
                    imports.add(foundImport);
                }
            }
        }
        scanner.close();
        javaReferenceFinder.populateReferences(analysis, content);
        return analysis;
    }
}