pmacs3/ctags.py

import os
import re
from stat import ST_MTIME
from subprocess import Popen, PIPE, STDOUT

class CtagsRunError(Exception): pass
class CtagsParseError(Exception): pass

# ctags --languages=... -R -excmd=number

class TagManager(object):
    lang  = None
    prune = set(['SCCS', 'RCS', 'CVS', '.svn', '.hg', '.git', '.bzr'])
    exts  = set()

    def __init__(self, base='.'):
        self.db = None
        self.base  = base
        self.path  = os.path.join(self.base, 'tags')
        self.update()

    def has(self, name):
        return name in self.db.tags

    def get(self, name):
        return self.db.tags.get(name, [])

    def update(self):
        if self.is_outdated():
            self.generate()
        self.db = Ctags(self.path)
        self.db.parse()

    def generate(self):
        args = ['ctags', '-f', self.path, '-L-', '--excmd=number']
        if self.lang:
            args.append('--language-force=%s' % self.lang)
            
        pipe = Popen(args, stdin=PIPE, stdout=PIPE, stderr=STDOUT)
        indata = '\n'.join(self.get_paths()) + '\n'
        outdata = pipe.communicate(indata)
        if pipe.returncode != 0:
            raise CtagsRunError("%r: %r" % (args, outdata))

    def get_paths(self):
        return list(self._walk())

    def is_outdated(self):
        if not os.path.exists(self.path):
            return True
        mtime = os.stat(self.path)[ST_MTIME]
        itr = self._walk(mtime)
        try:
            next(itr)
            return True
        except StopIteration:
            return False

    def _walk(self, mtime=0):
        paths = []
        for root, dirs, files in os.walk(self.base):
            for d in dirs:
                if d in self.prune:
                    dirs.remove(d)
            for f in files:
                path = os.path.join(root, f)
                if not self._match(path):
                    continue
                elif os.stat(path)[ST_MTIME] < mtime:
                    continue
                else:
                    yield os.path.join(root, f)
        return

    def _match(self, path):
        return os.path.splitext(path)[1] in self.exts

class Ctags(object):
    line_re = re.compile(r'^(.+?)\t(.+?)\t(.+?);"(.+)$')

    def __init__(self, path):
        self.path = path
        self.records = []
        self.tags = {}

    def get(self, tag):
        return self.tags.get(tag, [])

    def parse(self):
        for line in open(self.path, 'r'):
            if line.startswith('!'): continue

            m = self.line_re.match(line)
            if not m: raise CtagsParseError("%r" % line)

            name, path, ex, _ = m.groups()
            self.add(CtagRecord(name, path, ex))

    def add(self, r):
        self.records.append(r)
        self.tags.setdefault(r.name, [])
        self.tags[r.name].append(r)

class CtagRecord(object):
    def __init__(self, name, path, ex):
        if not ex.isdigit():
            raise Exception("excmd %r unsupported; use --excmd=number" % ex)
        self.name = name
        self.path = path
        self.line = int(ex)

    def __repr__(self):
        return "CtagRecord(%r, %r, %d)" % (self.name, self.path, self.line)

if __name__ == '__main__':
    import sys
    from pprint import pprint

    ctags = Ctags(sys.argv[1])
    ctags.parse()

    if len(sys.argv) > 2:
        print("records for %r:" % sys.argv[2])
        for r in ctags.get(sys.argv[2]):
            print(r)
    else:
        pprint(ctags.records)