def wordlist_to_graph_file(wordlist, dbfile, fieldname="_", strip=True):
"""Writes a word graph file from a list of words.
>>> # Open a word list file with one word on each line, and write the
>>> # word graph to a graph file
>>> wordlist_to_graph_file("mywords.txt", "mywords.dawg")
:param wordlist: an iterable containing the words for the graph. The words
must be in sorted order.
:param dbfile: a filename string or file-like object to write the word
graph to. This function will close the file.
"""
from whoosh.filedb.structfile import StructFile
if isinstance(dbfile, string_type):
dbfile = open(dbfile, "wb")
if not isinstance(dbfile, StructFile):
dbfile = StructFile(dbfile)
gw = dawg.GraphWriter(dbfile)
gw.start_field(fieldname)
for word in wordlist:
if strip:
word = word.strip()
gw.insert(word)
gw.finish_field()
gw.close()
def test_empty_key():
gw = dawg.GraphWriter(RamStorage().create_file("test"))
gw.start_field("_")
assert_raises(KeyError, gw.insert, b(""))
assert_raises(KeyError, gw.insert, "")
assert_raises(KeyError, gw.insert, u(""))
assert_raises(KeyError, gw.insert, [])
def gwrite(keys, st=None):
st = st or RamStorage()
f = st.create_file("test")
gw = dawg.GraphWriter(f)
gw.start_field("_")
for key in keys:
gw.insert(key)
gw.finish_field()
gw.close()
return st
def test_insert_bytes():
# This test is only meaningful on Python 3
domain = [b("alfa"), b("bravo"), b("charlie")]
st = RamStorage()
gw = dawg.GraphWriter(st.create_file("test"))
gw.start_field("test")
for key in domain:
gw.insert(key)
gw.close()
cur = dawg.GraphReader(st.open_file("test")).cursor()
assert_equal(list(cur.flatten()), domain)
def _fst_roundtrip(domain, t):
with TempStorage() as st:
f = st.create_file("test")
gw = dawg.GraphWriter(f, vtype=t)
gw.start_field("_")
for key, value in domain:
gw.insert(key, value)
gw.finish_field()
gw.close()
f = st.open_file("test")
gr = dawg.GraphReader(f, vtype=t)
cur = dawg.Cursor(gr)
assert_equal(list(cur.flatten_v()), domain)
f.close()
def test_insert_unicode():
domain = [
u("\u280b\u2817\u2801\u281d\u2809\u2811"),
u("\u65e5\u672c"),
u("\uc774\uc124\ud76c"),
]
st = RamStorage()
gw = dawg.GraphWriter(st.create_file("test"))
gw.start_field("test")
for key in domain:
gw.insert(key)
gw.close()
cur = dawg.GraphReader(st.open_file("test")).cursor()
assert_equal(list(cur.flatten_strings()), domain)
def test_within_unicode():
domain = [
u("\u280b\u2817\u2801\u281d\u2809\u2811"),
u("\u65e5\u672c"),
u("\uc774\uc124\ud76c"),
]
st = RamStorage()
gw = dawg.GraphWriter(st.create_file("test"))
gw.start_field("test")
for key in domain:
gw.insert(key)
gw.close()
gr = dawg.GraphReader(st.open_file("test"))
s = list(dawg.within(gr, u("\uc774.\ud76c")))
assert_equal(s, [u("\uc774\uc124\ud76c")])
def test_fields():
with TempStorage() as st:
f = st.create_file("test")
gw = dawg.GraphWriter(f)
gw.start_field("f1")
gw.insert("a")
gw.insert("aa")
gw.insert("ab")
gw.finish_field()
gw.start_field("f2")
gw.insert("ba")
gw.insert("baa")
gw.insert("bab")
gw.close()
gr = dawg.GraphReader(st.open_file("test"))
cur1 = dawg.Cursor(gr, gr.root("f1"))
cur2 = dawg.Cursor(gr, gr.root("f2"))
assert_equal(list(cur1.flatten_strings()), ["a", "aa", "ab"])
assert_equal(list(cur2.flatten_strings()), ["ba", "baa", "bab"])
gr.close()
def add_spelling(ix, fieldnames, commit=True):
"""Adds spelling files to an existing index that was created without
them, and modifies the schema so the given fields have the ``spelling``
attribute. Only works on filedb indexes.
>>> ix = index.open_dir("testindex")
>>> add_spelling(ix, ["content", "tags"])
:param ix: a :class:`whoosh.filedb.fileindex.FileIndex` object.
:param fieldnames: a list of field names to create word graphs for.
:param force: if True, overwrites existing word graph files. This is only
useful for debugging.
"""
from whoosh.filedb.filereading import SegmentReader
from whoosh.support import dawg
writer = ix.writer()
storage = writer.storage
schema = writer.schema
segments = writer.segments
for segment in segments:
r = SegmentReader(storage, schema, segment)
f = segment.create_file(storage, ".dag")
gw = dawg.GraphWriter(f)
for fieldname in fieldnames:
gw.start_field(fieldname)
for word in r.lexicon(fieldname):
gw.insert(word)
gw.finish_field()
gw.close()
for fieldname in fieldnames:
schema[fieldname].spelling = True
if commit:
writer.commit(merge=False)
def test_keys_out_of_order():
f = RamStorage().create_file("test")
gw = dawg.GraphWriter(f)
gw.start_field("test")
gw.insert("alfa")
assert_raises(KeyError, gw.insert, "abba")
def test_empty_fieldname():
gw = dawg.GraphWriter(RamStorage().create_file("test"))
assert_raises(ValueError, gw.start_field, "")
assert_raises(ValueError, gw.start_field, None)
assert_raises(ValueError, gw.start_field, 0)
