Programmatic access to Java documentation [closed]

Question

Is there an api to programmatically access to the Java documentation? Something similar to what editors/IDEs like Eclipse do for IntelliSense. Surely, the javadoc generated html can be parsed and indexed, but would be great to know if something already exists as a standalone package.

Update: To be clear, I am trying to get a programmatic access to the documentation for the java language implementation. However, there isn't an easy way out, but you have to get the openjdk from Oracle's site and then do make docs and supply some doclet to get the output in non-html form. This is what I was trying to avoid from the beginning. So I'm reading all the Makefiles now trying to figure out why $PLATFORM isn't being resolved and source build is failing!

What I finally settled with:

Scraping html docs. Yep. Realized that'd get my job done.

import urllib2
import pymongo

mongo = pymongo.Connection('localhost')
mongo_db = mongo['api_db']
mongo_collection = mongo_db['api_collection']

URL='''http://docs.oracle.com/javase/6/docs/api/index-files/index-'''
URL_range = 27+1

def getPageElements(url):
    content=urllib2.urlopen(url).read()
    from BeautifulSoup import BeautifulSoup
    soup = BeautifulSoup(content)
    elements = soup('dl')
    return elements


def savePageElements(elements):
    for i in elements[0]:
            try:
                    entry={'type_description':str(i.findNext('dt')).split('-')[1].split('<')[0]\
                                    +i.findNext('a').findNext('a').renderContents(),\
                                    'signature':i.findNext('a').findNext('a').renderContents()+i.findNext('b').renderContents(),\
                                    'description': i.findNext('dd').renderContents()\
                            }   
                    print entry
                    insert_id = mongo_collection.save(entry)
            except:
                    pass


def retrieve(str):
     mongo_documents = mongo_collection.find({
    'signature': str,
    #'type_description':"Method of java.io.PrintStream"
    })  
    for this_document in mongo_documents:
        print this_document


if __name__=="__main__":
    for i in range(1,URL_range):
            #url=URL+str(i)+".html"
            #print "Processing:", url
            #elements=getPageElements(url)
            #print elements[0]
            #savePageElements(elements)
            retrieve("println(String)")

But take a look at dexy. If I could have managed to build OpenJDK on Ubuntu without issue - it would have generated nice JSON to play with.

Answer 1

As you write Eclipse does parse javadoc and does pretty good job of it. It seems it was introduced in 3.2M4. You may be able to extract the functionality from Eclipse sources, checking with Eclipse community for entry point may be a good place to start.

If that does not work, I guess your starting point would be a plain vanilla HTML parser.