-
Notifications
You must be signed in to change notification settings - Fork 0
/
app.py
227 lines (167 loc) · 6.46 KB
/
app.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
import logging
import json
from collections import defaultdict
from datetime import datetime, timedelta
import boto3
import weaviate
from langchain.docstore.document import Document
from langchain.chains.summarize import load_summarize_chain
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.llms import OpenAI
logger = logging.getLogger()
DATE = (datetime.now() - timedelta(days=1)).strftime("%Y-%m-%d")
ARTICLE_LIMIT = 3
topics = ["business", "entertainment", "nation", "science", "technology", "world"]
config = {}
def news_by_topic(topic: str):
"""Get all articles from Weaviate by topic"""
results = []
offset = 0
client = weaviate.Client(
url=config["WEAVIATE_URL"],
auth_client_secret=weaviate.AuthApiKey(api_key=config["WEAVIATE_API_KEY"])
)
REQ_LIMIT = 100
if ARTICLE_LIMIT < REQ_LIMIT:
REQ_LIMIT = ARTICLE_LIMIT
where_filter = {
"operator": "And",
"operands": [{
"path": ["published_date"],
"operator": "Equal",
"valueString": DATE
},{
"path": ["topic"],
"operator": "Equal",
"valueString": topic
}]
}
while True:
result = (
client.query
.get("Article", ["title", "text", "url"])
.with_limit(REQ_LIMIT)
.with_offset(offset)
.with_additional(["vector"])
.with_where(where_filter)
.do()
)
result = result['data']['Get']['Article']
results.extend(result)
if len(result) < REQ_LIMIT or len(results) >= ARTICLE_LIMIT:
break
offset += REQ_LIMIT
return results
def summarize_article(article_text: str):
"""Summarize article using OpenAI API & Langchain"""
article_doc = Document(page_content=article_text, metadata={"source": str(0)})
text_splitter = RecursiveCharacterTextSplitter(
# Set a really small chunk size, just to show.
chunk_size = 1000,
chunk_overlap = 20,
)
texts = text_splitter.split_documents([article_doc])
llm = OpenAI(temperature=0, openai_api_key=config["OPENAI_API_KEY"],
model_name=config["OPENAI_MODEL_NAME"])
chain = load_summarize_chain(llm, chain_type="map_reduce")
return chain.run(texts)
def get_summarized_articles() -> defaultdict:
"""Get summarized articles from Weaviate by topic"""
summarized_articles = defaultdict(dict)
for topic in topics:
articles = news_by_topic(topic)
logging.debug(topic)
for article in articles:
# TODO: Take care of in ETL
if "text" not in article or not article["text"]:
logging.debug("Skipping {}, no text".format(article["title"]))
continue
if article["title"] in summarized_articles[topic]:
logging.debug("Skipping {}, already in dict".format(article["title"]))
continue
article["text"] = summarize_article(article["text"])
summarized_articles[topic][article["title"]] = article
return summarized_articles
def write_summ_article(article: dict, html: str) -> str:
"""Write summarized article to HTML"""
logging.debug("Starting: write_summ_article for {}".format(article["title"]))
article_html = f"""<h3>{article["title"].replace('.json', '')}</h3>
<p>{article["text"]}</p>
<a href=\"{article["url"]}\"> source </a>
"""
html += article_html
logging.debug("Finished: write_summ_article for {}".format(article["title"]))
return html
def write_section(articles: list[dict], topic: str):
"""Write topic section to HTML"""
logging.debug(f"Starting: write_section for {topic} with {len(articles)} articles")
html=f"""<h1>{topic.capitalize()}</h1>\n<hr>\n"""
for article in articles.values():
html = write_summ_article(article, html)
html += "\n"
logging.debug(f"Finished: write_section for {topic}")
return html
def write_html():
logging.debug("Staring: write_html")
"""Write HTML file"""
summarized_articles = get_summarized_articles()
article_html = f"""<!DOCTYPE html>
<html lang="en-US">
<head>
<meta charset="UTF-8">
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>NeuralDigest Summary: {DATE}</title>
<link href="https://fonts.googleapis.com/css2?family=Roboto:wght@400;700&display=swap" rel="stylesheet">
<link href="../style.css" rel="stylesheet">
</head>
<h1> Summary for News Articles: {DATE} </h1>
<hr>
<body>
"""
for topic in topics:
article_html += write_section(summarized_articles[topic], topic)
article_html += "</html>"
logging.debug("Finished: write_html")
return article_html
def lambda_handler(event, _):
if "LOG_LEVEL" in event and event["LOG_LEVEL"].lower() in [
"debug",
"info",
"warning",
"error",
"critical",
]:
logger.setLevel(level=getattr(logging, event["LOG_LEVEL"].upper()))
else:
logger.setLevel(level=logging.INFO)
logging.debug("Initiating lambda_handler")
s3 = boto3.client('s3')
file_name = f'articles/{DATE}.html'
bucket_name = event["S3_BUCKET_NAME"]
configVars = [
"WEAVIATE_URL",
"WEAVIATE_API_KEY",
"OPENAI_API_KEY",
"OPENAI_MODEL_NAME"
]
for conf in configVars:
try:
config[conf] = event[conf]
except KeyError:
logging.warning(f"Missing event variable: {conf}")
try:
# Write article_html to S3 bucket
s3.put_object(Body=write_html(), Bucket=bucket_name, Key=file_name, ContentType='text/html')
logging.info(f'Successfully written {file_name} to {bucket_name}')
except Exception as e:
logging.warning(f'Error writing {file_name} to {bucket_name}: {str(e)}')
logging.debug("Retrieving and modifying articles.json")
articles_json_key = "articles.json"
response = s3.get_object(Bucket=bucket_name, Key=articles_json_key)
content = response['Body'].read().decode('utf-8')
articles = json.loads(content)
if DATE not in articles:
articles.insert(0, DATE)
modified_content = json.dumps(articles)
s3.put_object(Bucket=bucket_name, Key=articles_json_key, Body=modified_content)