Simple library for building and searching an index from a subversion repository

svnsearch.py

import pysvn
import datetime
import os
from whoosh.index import create_in, open_dir
from whoosh.fields import *
from whoosh.qparser import QueryParser

schema = Schema(author=TEXT(stored=True), 
        message=TEXT(stored=True),
        revision=KEYWORD(stored=True),
        timestamp=DATETIME(stored=True))

def open_index(fname):
    ''' opens an index at the path using the specified file name '''
    if not os.path.exists(fname):
        os.makedirs(fname)
    index = create_in(fname, schema)
    return index

def get_svn_login(realm, username, may_save):
    ''' prompts user for svn login if needed '''
    username = raw_input('Username: ')
    password = raw_input('Password: ')
    return (True, username, password, True)

def index_svn_commits(svn_url, index_path):
    ''' extracts commit messages from subversion at the given URL and indexes the commit messages into the given index path'''
    c = pysvn.Client()
    c.callback_get_login = get_svn_login
    msgs = c.log(svn_url)
    index = open_index(index_path)
    writer = index.writer()
    for msg in msgs:
        writer.add_document(author = unicode(msg.author),
                message = unicode(msg.message),
                revision = unicode(msg.revision.number),
                timestamp = datetime.datetime.fromtimestamp(msg.date))
    writer.commit(optimize=True)

def dump_terms(index_path, field):
    ''' dumps the list of terms in the given field '''
    index = open_dir(index_path)
    search = index.searcher() 
    terms = list(searcher.lexicon(field))
    return terms

def search_commits(index_path, query):
    ''' searches the index using the given query object and returns matching commits '''
    index = open_dir(index_path)
    searcher = index.searcher() 
    results = searcher.search(query)
    return results

def get_query(params):
    ''' takes a dictionary and constructs a whoosh query object for searching '''
    qp = QueryParser('message', schema)
    q_phrases = []
    for field,value in params.items():
        q_phrases.append('%s:%s' % (field, unicode(value)))
    return qp.parse(' AND '.join(q_phrases))