From 72f6bb2fd31e5dad58314090f02607693265657d Mon Sep 17 00:00:00 2001
From: Robert Altner <robert-altner@t-online.de>
Date: Fri, 10 Jan 2020 20:23:22 +0100
Subject: [PATCH] improved smart
---
api/thesaurus.py | 28 +++++++++++++++++++++++++---
1 file changed, 25 insertions(+), 3 deletions(-)
diff --git a/api/thesaurus.py b/api/thesaurus.py
index e240b05..9a60dfd 100644
--- a/api/thesaurus.py
+++ b/api/thesaurus.py
@@ -1,12 +1,14 @@
import requests
import re, random
+import nltk
from util import logging, config
key = config.settings["api_keys"]["thesaurus"]
url = f"http://thesaurus.altervista.org/thesaurus/v1?key={key}&language=en_US&output=json&word="
def thesaurufy_sentence(sentence):
- symbols = re.findall(r"[\w']+|[.,!?;]", sentence)
+ symbols = nltk.word_tokenize(sentence)
+ tags = nltk.pos_tag(symbols)
if len(symbols) == 0:
return ""
@@ -15,6 +17,10 @@ def thesaurufy_sentence(sentence):
if not symbols[i].isalpha():
continue
+ if tags[i][1] != "NN" and tags[i][1] != "VB" and tags[i][1] != "VBG" and tags[i][1] != "VBP" and tags[i][1] != "JJ" and tags[i][1] != "RB":
+ symbols[i] = " " + symbols[i]
+ continue
+
response = requests.get(url + symbols[i])
if not response.ok:
# logging.warning(f"Thesaurus API returned {response.status_code} ({url + symbols[i]})")
@@ -22,14 +28,30 @@ def thesaurufy_sentence(sentence):
continue
responses = response.json()["response"]
- words = responses[random.randint(0, len(responses) - 1)]["list"]["synonyms"].split("|")
+ condition = ""
+ if tags[i][1] == "JJ":
+ condition = "(adj)"
+ elif tags[i][1] == "RB":
+ condition = "(adv)"
+ elif tags[i][1] == "NN":
+ condition = "(noun)"
+ else:
+ condition = "(verb)"
+
+ try:
+ words = random.choice([a for a in responses if a["list"]["category"] == condition])["list"]["synonyms"].split("|")
+ except:
+ symbols[i] = " " + symbols[i]
+ continue
+
+ # print(words)
word = words[random.randint(0, len(words) - 1)]
if "(" in word:
if "antonym" in word.split("(")[1].lower():
symbols[i] = " " + symbols[i]
continue
-
+
word = word.split("(")[0][:-1]
symbols[i] = " " + word