try:
import apt
import apt_pkg
HAS_APT=True
except ImportError:
HAS_APT=False
import xapian
import re
import os, os.path

class Descriptions:
def info(self):
"""
Return general information about the plugin.

The information returned is a dict with various keywords:

timestamp (required)
the last modified timestamp of this data source. This will be used
to see if we need to update the database or not. A timestamp of 0
means that this data source is either missing or always up to date.
values (optional)
an array of dicts { name: name, desc: description }, one for every
numeric value indexed by this data source.

Note that this method can be called before init. The idea is that, if
the timestamp shows that this plugin is currently not needed, then the
long initialisation can just be skipped.
"""
res = dict(
timestamp=0,
prefixes=[
dict(idx="Z", qp=None, type=None,
desc="Stemmed forms of keywords",
ldesc="This contains the stemmed forms of keywords as generated by"
" TermGenerator and matched by QueryParser"),
],
)
if not HAS_APT: return res
if not hasattr(apt_pkg, "config"): return res
fname = apt_pkg.config.find_file("Dir::Cache::pkgcache")
if not os.path.exists(fname): return res
res["sources"] = [dict(path=fname, desc="APT index")]
res["timestamp"] = os.path.getmtime(fname)
return res

def init(self, info, progress):
"""
If needed, perform long initialisation tasks here.

info is a dictionary with useful information. Currently it contains
the following values:

"values": a dict mapping index mnemonics to index numbers

The progress indicator can be used to report progress.
"""
self.stemmer = xapian.Stem("english")
self.indexer = xapian.TermGenerator()
self.indexer.set_stemmer(self.stemmer)

def send_extra_info(self, db=None, **kw):
"""
Receive extra parameters from the indexer.

This may be called more than once, but after init().

We are using this to get the database instance
"""
if db is not None:
self.indexer.set_flags(xapian.TermGenerator.FLAG_SPELLING)
self.indexer.set_database(db)

def doc(self):
"""
Return documentation information for this data source.

The documentation information is a dictionary with these keys:
name: the name for this data source
shortDesc: a short description
fullDoc: the full description as a chapter in ReST format
"""
return dict(
name = "Package descriptions",
shortDesc = "terms extracted from the package descriptions using Xapian's TermGenerator",
fullDoc = """
The Descriptions data source simply uses Xapian's TermGenerator to
tokenise and index the package descriptions.

Currently this creates normal terms as well as stemmed terms
prefixed with ``Z``.
"""
)

def index(self, document, pkg):
"""
Update the document with the information from this data source.

document is the document to update
pkg is the python-apt Package object for this package
"""
self.indexer.set_document(document)

# Index the record
self.indexer.index_text_without_positions(pkg.name)
version = pkg.candidate
if version is not None:
self.indexer.index_text_without_positions(version.raw_description)
if not HAS_APT:
def index(self, document, pkg):
pass

def indexDeb822(self, document, pkg):
"""
Update the document with the information from this data source.

This is alternative to index, and it is used when indexing with package
data taken from a custom Packages file.

document is the document to update
pkg is the Deb822 object for this package
"""
self.indexer.set_document(document)

# Index the record
self.indexer.index_text_without_positions(pkg["Package"])

if 'Description' in pkg:
self.indexer.index_text_without_positions(pkg["Description"])
else:
# check if we have a translated description
for k in pkg.keys():
if k.startswith('Description-'):
self.indexer.index_text_without_positions(pkg[k])
break

def init(**kw):
"""
Create and return the plugin object.
"""
return Descriptions()