Update scripture_compare.py
Browse files- scripture_compare.py +32 -29
scripture_compare.py
CHANGED
@@ -1,12 +1,11 @@
|
|
1 |
import gradio as gr
|
2 |
import urllib.request, json
|
3 |
-
from transformers.utils import logging
|
4 |
-
logging.set_verbosity_error()
|
5 |
-
|
6 |
from sentence_transformers import SentenceTransformer
|
7 |
from sentence_transformers import util
|
|
|
|
|
8 |
|
9 |
-
# Search URL for
|
10 |
SEARCH_URL = 'https://dd4-biblical.appspot.com/_api/scriptures/v1/search?searchText={}&lang=en&version={}'
|
11 |
|
12 |
model = SentenceTransformer("all-MiniLM-L6-v2")
|
@@ -17,18 +16,18 @@ class Candidate:
|
|
17 |
compareText = ''
|
18 |
score = 0
|
19 |
|
20 |
-
def __init__(self, book, chapter, verse,
|
21 |
self.book = book
|
22 |
self.chapter = chapter
|
23 |
self.verse = verse
|
24 |
-
self.standardText =
|
25 |
|
26 |
def reference(self):
|
27 |
return '{} {}:{}'.format(self.book, self.chapter, self.verse)
|
28 |
|
29 |
def to_tabbed(self):
|
30 |
return '{}\t\t {:.4f}\t\t {}\t\t {}'.format(self.reference(), self.score, self.standardText, self.compareText)
|
31 |
-
|
32 |
def to_csv(self):
|
33 |
return '{},{:.4f},"{}","{}"'.format(self.reference(), self.score, self.standardText, self.compareText)
|
34 |
|
@@ -38,11 +37,12 @@ class Candidate:
|
|
38 |
def __repr__(self):
|
39 |
return self.to_csv()
|
40 |
|
41 |
-
|
|
|
42 |
candidates = []
|
43 |
-
|
44 |
# Connect to the api and get the standard translation
|
45 |
-
with urllib.request.urlopen(SEARCH_URL.format(urllib.parse.quote(reference),
|
46 |
response = json.load(url)
|
47 |
# print(response)
|
48 |
|
@@ -50,34 +50,34 @@ def compare(reference, standardVersion, compareVersion):
|
|
50 |
# print(result)
|
51 |
candidate = Candidate(result['book'], result['chapter'], result['verse'], result['text'])
|
52 |
candidates.append(candidate)
|
53 |
-
|
54 |
-
|
55 |
# Then fetch the comparison text
|
56 |
-
with urllib.request.urlopen(SEARCH_URL.format(urllib.parse.quote(reference),
|
57 |
response = json.load(url)
|
58 |
# print(response)
|
59 |
|
60 |
for result in response['items']:
|
61 |
# print(result)
|
62 |
-
candidate =
|
63 |
-
if candidate is None
|
64 |
candidate = Candidate(result['book'], result['chapter'], result['verse'], '')
|
65 |
candidates.append(candidate)
|
66 |
-
|
67 |
|
68 |
candidate.compareText = result['text']
|
69 |
|
70 |
# Isa 1:1 standardText: This is the book of Isaiah compareText: The book that Isaiah wrote.
|
71 |
# Isa 1:2 standardText: Isaiah was a good man compareText: Isaiah did what was right.
|
72 |
-
|
73 |
-
|
74 |
-
|
75 |
for candidate in candidates:
|
76 |
-
|
77 |
-
|
78 |
-
|
79 |
-
embeddings1 = model.encode(
|
80 |
-
embeddings2 = model.encode(
|
81 |
cosine_scores = util.cos_sim(embeddings1, embeddings2)
|
82 |
|
83 |
for i in range(len(candidates)):
|
@@ -87,16 +87,19 @@ def compare(reference, standardVersion, compareVersion):
|
|
87 |
return '\n\n'.join(c.to_tabbed() for c in candidates)
|
88 |
|
89 |
|
90 |
-
|
91 |
compare,
|
92 |
[
|
93 |
gr.Textbox(
|
94 |
label="Scriptures",
|
95 |
-
info="Enter a verse, verse range, chapter(s) or mix
|
|
|
96 |
lines=1,
|
97 |
),
|
98 |
-
gr.Radio(["ISR", "RSKJ", "NRSV", "NWT", "SEP", "DSS"], label="Control Version",
|
99 |
-
|
|
|
|
|
100 |
],
|
101 |
gr.Textbox(label="Comparisons", lines=28))
|
102 |
-
|
|
|
1 |
import gradio as gr
|
2 |
import urllib.request, json
|
|
|
|
|
|
|
3 |
from sentence_transformers import SentenceTransformer
|
4 |
from sentence_transformers import util
|
5 |
+
from transformers.utils import logging
|
6 |
+
logging.set_verbosity_error()
|
7 |
|
8 |
+
# Search URL for Maccabees Ministries scriptures. Will fetch scriptures as a json object.
|
9 |
SEARCH_URL = 'https://dd4-biblical.appspot.com/_api/scriptures/v1/search?searchText={}&lang=en&version={}'
|
10 |
|
11 |
model = SentenceTransformer("all-MiniLM-L6-v2")
|
|
|
16 |
compareText = ''
|
17 |
score = 0
|
18 |
|
19 |
+
def __init__(self, book, chapter, verse, standard_text):
|
20 |
self.book = book
|
21 |
self.chapter = chapter
|
22 |
self.verse = verse
|
23 |
+
self.standardText = standard_text
|
24 |
|
25 |
def reference(self):
|
26 |
return '{} {}:{}'.format(self.book, self.chapter, self.verse)
|
27 |
|
28 |
def to_tabbed(self):
|
29 |
return '{}\t\t {:.4f}\t\t {}\t\t {}'.format(self.reference(), self.score, self.standardText, self.compareText)
|
30 |
+
|
31 |
def to_csv(self):
|
32 |
return '{},{:.4f},"{}","{}"'.format(self.reference(), self.score, self.standardText, self.compareText)
|
33 |
|
|
|
37 |
def __repr__(self):
|
38 |
return self.to_csv()
|
39 |
|
40 |
+
|
41 |
+
def compare(reference, standard_version, compare_version):
|
42 |
candidates = []
|
43 |
+
candidate_map = {}
|
44 |
# Connect to the api and get the standard translation
|
45 |
+
with urllib.request.urlopen(SEARCH_URL.format(urllib.parse.quote(reference), standard_version)) as url:
|
46 |
response = json.load(url)
|
47 |
# print(response)
|
48 |
|
|
|
50 |
# print(result)
|
51 |
candidate = Candidate(result['book'], result['chapter'], result['verse'], result['text'])
|
52 |
candidates.append(candidate)
|
53 |
+
candidate_map[candidate.reference()] = candidate
|
54 |
+
|
55 |
# Then fetch the comparison text
|
56 |
+
with urllib.request.urlopen(SEARCH_URL.format(urllib.parse.quote(reference), compare_version)) as url:
|
57 |
response = json.load(url)
|
58 |
# print(response)
|
59 |
|
60 |
for result in response['items']:
|
61 |
# print(result)
|
62 |
+
candidate = candidate_map['{} {}:{}'.format(result['book'], result['chapter'], result['verse'])]
|
63 |
+
if candidate is None:
|
64 |
candidate = Candidate(result['book'], result['chapter'], result['verse'], '')
|
65 |
candidates.append(candidate)
|
66 |
+
candidate_map[candidate.reference()] = candidate
|
67 |
|
68 |
candidate.compareText = result['text']
|
69 |
|
70 |
# Isa 1:1 standardText: This is the book of Isaiah compareText: The book that Isaiah wrote.
|
71 |
# Isa 1:2 standardText: Isaiah was a good man compareText: Isaiah did what was right.
|
72 |
+
|
73 |
+
standard_texts = [] # 2 items This is the book of Isaiah, Isaiah was a good man
|
74 |
+
compare_texts = [] # 2 items The book that Isaiah wrote., Isaiah did what was right.
|
75 |
for candidate in candidates:
|
76 |
+
standard_texts.append(candidate.standardText)
|
77 |
+
compare_texts.append(candidate.compareText)
|
78 |
+
|
79 |
+
embeddings1 = model.encode(standard_texts, convert_to_tensor=True)
|
80 |
+
embeddings2 = model.encode(compare_texts, convert_to_tensor=True)
|
81 |
cosine_scores = util.cos_sim(embeddings1, embeddings2)
|
82 |
|
83 |
for i in range(len(candidates)):
|
|
|
87 |
return '\n\n'.join(c.to_tabbed() for c in candidates)
|
88 |
|
89 |
|
90 |
+
ui = gr.Interface(
|
91 |
compare,
|
92 |
[
|
93 |
gr.Textbox(
|
94 |
label="Scriptures",
|
95 |
+
info="Enter a verse, verse range, chapter(s) or mix "
|
96 |
+
+ "(e.g. Gen 2:3 or Gen 2:1-3 or Exo 12 or Lev 20-23 or Gen 2:3,Exo 12,Lev 20-23)",
|
97 |
lines=1,
|
98 |
),
|
99 |
+
gr.Radio(["ISR", "RSKJ", "NRSV", "NWT", "SEP", "DSS"], label="Control Version",
|
100 |
+
info="The version of scripture to use as the standard.", value="RSKJ"),
|
101 |
+
gr.Radio(["ISR", "RSKJ", "NRSV", "NWT", "SEP", "DSS"], label="Comparison Version",
|
102 |
+
info="The version of scripture to use as the comparison.", value="SEP")
|
103 |
],
|
104 |
gr.Textbox(label="Comparisons", lines=28))
|
105 |
+
ui.launch()
|