diff options
author | Martin Czygan <martin.czygan@gmail.com> | 2020-11-12 00:19:30 +0100 |
---|---|---|
committer | Martin Czygan <martin.czygan@gmail.com> | 2020-11-12 00:19:30 +0100 |
commit | f0b5857b45b0982cb587fce6b847d1d039794636 (patch) | |
tree | ed55bee41027df04e04a9ffeecef6e3b1c92ec2f /fuzzycat | |
parent | 001e1b99a126d67a9c65b2995bb2f3873721a36b (diff) | |
download | fuzzycat-f0b5857b45b0982cb587fce6b847d1d039794636.tar.gz fuzzycat-f0b5857b45b0982cb587fce6b847d1d039794636.zip |
update deps
Diffstat (limited to 'fuzzycat')
-rw-r--r-- | fuzzycat/build.py | 2 | ||||
-rw-r--r-- | fuzzycat/cluster.py | 4 | ||||
-rw-r--r-- | fuzzycat/main.py | 11 |
3 files changed, 10 insertions, 7 deletions
diff --git a/fuzzycat/build.py b/fuzzycat/build.py index e49e7d7..49a061f 100644 --- a/fuzzycat/build.py +++ b/fuzzycat/build.py @@ -3,13 +3,13 @@ WIP: Build auxiliary data structures for lookup. """ import fileinput +import json import operator import sqlite3 import string import sys import tempfile -import orjson as json from nltk import word_tokenize from nltk.corpus import stopwords diff --git a/fuzzycat/cluster.py b/fuzzycat/cluster.py index 2005c2f..dd55a24 100644 --- a/fuzzycat/cluster.py +++ b/fuzzycat/cluster.py @@ -69,10 +69,10 @@ import string import subprocess import sys import tempfile +from dataclasses import dataclass, field from typing import Any, Callable, Dict, Generator, List, Optional, Tuple import fuzzy -from dataclasses import dataclass, field __all__ = [ "release_key_title", @@ -95,6 +95,7 @@ class Contrib: surname: Optional[str] role: Optional[str] + @dataclass class KeyDoc: """ @@ -104,6 +105,7 @@ class KeyDoc: title: str contribs: List[Contrib] = field(default_factory=list) + @dataclass class ClusterResult: """ diff --git a/fuzzycat/main.py b/fuzzycat/main.py index 6086798..900d5c0 100644 --- a/fuzzycat/main.py +++ b/fuzzycat/main.py @@ -12,16 +12,18 @@ Run, e.g. fuzzycat cluster --help for more options. Example: import argparse import cProfile as profile +import fileinput +import json import io import logging import pstats import sys import tempfile -import fileinput from fuzzycat.build import NgramLookup, TitleTokenList -from fuzzycat.cluster import (Cluster, release_key_title, release_key_title_normalized, - release_key_title_nysiis, release_key_title_ngram) +from fuzzycat.cluster import (Cluster, release_key_title, release_key_title_ngram, + release_key_title_normalized, release_key_title_nysiis) + def run_cluster(args): logger = logging.getLogger('main.run_cluster') @@ -44,8 +46,7 @@ def run_verify(args): TODO. Ok, we should not fetch data we have on disk (at the clustering step). """ - for line in fileinput.input(files=args.files): - pass + pass def run_build(args): |