Update app.py
Browse files
app.py
CHANGED
|
@@ -1,58 +1,205 @@
|
|
| 1 |
import gradio as gr
|
| 2 |
import re
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
|
| 4 |
-
# Функция для анализа поста
|
| 5 |
def analyze_post(content_type, link_to_post, post_likes, post_date, description, comment_count, all_comments):
|
| 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 |
# Создаем интерфейс Gradio
|
| 36 |
iface = gr.Interface(
|
| 37 |
fn=analyze_post,
|
| 38 |
inputs=[
|
| 39 |
-
gr.Radio(
|
| 40 |
-
|
| 41 |
-
|
| 42 |
-
|
| 43 |
-
|
| 44 |
-
gr.
|
| 45 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 46 |
],
|
| 47 |
outputs=[
|
| 48 |
-
gr.Textbox(label="
|
| 49 |
-
gr.Textbox(label="Usernames (Output 1)"),
|
| 50 |
-
gr.Textbox(label="Comments (Output 2)"),
|
| 51 |
-
gr.Textbox(label="Likes Chronology (Output 3)"),
|
| 52 |
-
gr.Textbox(label="Total Likes on Comments (Output 4)")
|
| 53 |
],
|
| 54 |
-
title="Instagram Comment Analyzer",
|
| 55 |
-
description="
|
| 56 |
)
|
| 57 |
|
| 58 |
-
|
|
|
|
|
|
| 1 |
import gradio as gr
|
| 2 |
import re
|
| 3 |
+
from collections import Counter
|
| 4 |
+
from datetime import datetime
|
| 5 |
+
import emoji
|
| 6 |
+
|
| 7 |
+
def clean_text(text):
|
| 8 |
+
"""Очищает текст от лишних пробелов и переносов строк"""
|
| 9 |
+
return ' '.join(text.split())
|
| 10 |
+
|
| 11 |
+
def count_emojis(text):
|
| 12 |
+
"""Подсчитывает количество эмодзи в тексте"""
|
| 13 |
+
return len([c for c in text if c in emoji.EMOJI_DATA])
|
| 14 |
+
|
| 15 |
+
def extract_mentions(text):
|
| 16 |
+
"""Извлекает упоминания пользователей из текста"""
|
| 17 |
+
return re.findall(r'@[\w\.]+', text)
|
| 18 |
+
|
| 19 |
+
def analyze_sentiment(text):
|
| 20 |
+
"""Простой анализ тональности по эмодзи и ключевым словам"""
|
| 21 |
+
positive_indicators = ['🔥', '❤️', '👍', 'круто', 'супер', 'класс', 'огонь', 'пушка']
|
| 22 |
+
negative_indicators = ['👎', '😢', 'плохо', 'ужас']
|
| 23 |
+
|
| 24 |
+
text_lower = text.lower()
|
| 25 |
+
positive_count = sum(1 for ind in positive_indicators if ind in text_lower)
|
| 26 |
+
negative_count = sum(1 for ind in negative_indicators if ind in text_lower)
|
| 27 |
+
|
| 28 |
+
if positive_count > negative_count:
|
| 29 |
+
return 'positive'
|
| 30 |
+
elif negative_count > positive_count:
|
| 31 |
+
return 'negative'
|
| 32 |
+
return 'neutral'
|
| 33 |
+
|
| 34 |
+
def extract_comment_data(comment_text):
|
| 35 |
+
"""
|
| 36 |
+
Извлекает данные из отдельного комментария
|
| 37 |
+
Возвращает (username, comment_text, likes_count, week_number)
|
| 38 |
+
"""
|
| 39 |
+
# Извлекаем имя пользователя
|
| 40 |
+
username_match = re.search(r"Фото профиля ([^\n]+)", comment_text)
|
| 41 |
+
if not username_match:
|
| 42 |
+
return None, None, 0, 0
|
| 43 |
+
|
| 44 |
+
username = username_match.group(1).strip()
|
| 45 |
+
|
| 46 |
+
# Извлекаем текст комментария (теперь без имени пользователя)
|
| 47 |
+
comment_pattern = fr"{username}\n(.*?)(?:\d+ нед\.)"
|
| 48 |
+
comment_match = re.search(comment_pattern, comment_text, re.DOTALL)
|
| 49 |
+
if comment_match:
|
| 50 |
+
# Очищаем текст комментария от упоминаний пользователя в начале
|
| 51 |
+
comment = clean_text(comment_match.group(1))
|
| 52 |
+
comment = re.sub(fr'^{username}\s*', '', comment)
|
| 53 |
+
comment = re.sub(r'^@[\w\.]+ ', '', comment) # Удаляем упоминания в начале
|
| 54 |
+
else:
|
| 55 |
+
comment = ""
|
| 56 |
+
|
| 57 |
+
# Извлекаем количество недель
|
| 58 |
+
week_match = re.search(r'(\d+) нед\.', comment_text)
|
| 59 |
+
weeks = int(week_match.group(1)) if week_match else 0
|
| 60 |
+
|
| 61 |
+
# Ищем количество лайков
|
| 62 |
+
likes = 0
|
| 63 |
+
likes_patterns = [
|
| 64 |
+
r"(\d+) отметк[аи] \"Нравится\"",
|
| 65 |
+
r"Нравится: (\d+)",
|
| 66 |
+
]
|
| 67 |
+
|
| 68 |
+
for pattern in likes_patterns:
|
| 69 |
+
likes_match = re.search(pattern, comment_text)
|
| 70 |
+
if likes_match:
|
| 71 |
+
likes = int(likes_match.group(1))
|
| 72 |
+
break
|
| 73 |
+
|
| 74 |
+
return username, comment.strip(), likes, weeks
|
| 75 |
|
|
|
|
| 76 |
def analyze_post(content_type, link_to_post, post_likes, post_date, description, comment_count, all_comments):
|
| 77 |
+
try:
|
| 78 |
+
# Разделяем комментарии по "Фото профиля"
|
| 79 |
+
comments_blocks = re.split(r'(?=Фото профиля)', all_comments)
|
| 80 |
+
comments_blocks = [block for block in comments_blocks if block.strip()]
|
| 81 |
+
|
| 82 |
+
# Основные списки для данных
|
| 83 |
+
usernames = []
|
| 84 |
+
comments = []
|
| 85 |
+
likes = []
|
| 86 |
+
weeks = []
|
| 87 |
+
|
| 88 |
+
# Дополнительные метрики
|
| 89 |
+
total_emojis = 0
|
| 90 |
+
mentions = []
|
| 91 |
+
sentiments = []
|
| 92 |
+
comment_lengths = []
|
| 93 |
+
|
| 94 |
+
# Обработка каждого комментария
|
| 95 |
+
for block in comments_blocks:
|
| 96 |
+
username, comment, like_count, week_number = extract_comment_data(block)
|
| 97 |
+
if username and comment:
|
| 98 |
+
usernames.append(username)
|
| 99 |
+
comments.append(comment)
|
| 100 |
+
likes.append(str(like_count))
|
| 101 |
+
weeks.append(week_number)
|
| 102 |
+
|
| 103 |
+
# Сбор дополнительных метрик
|
| 104 |
+
total_emojis += count_emojis(comment)
|
| 105 |
+
mentions.extend(extract_mentions(comment))
|
| 106 |
+
sentiments.append(analyze_sentiment(comment))
|
| 107 |
+
comment_lengths.append(len(comment))
|
| 108 |
+
|
| 109 |
+
# Аналитика
|
| 110 |
+
total_comments = len(comments)
|
| 111 |
+
avg_comment_length = sum(comment_lengths) / total_comments if comment_lengths else 0
|
| 112 |
+
sentiment_distribution = Counter(sentiments)
|
| 113 |
+
most_active_users = Counter(usernames).most_common(5)
|
| 114 |
+
most_mentioned = Counter(mentions).most_common(5)
|
| 115 |
+
avg_likes = sum(map(int, likes)) / len(likes) if likes else 0
|
| 116 |
+
earliest_week = max(weeks) if weeks else 0
|
| 117 |
+
latest_week = min(weeks) if weeks else 0
|
| 118 |
+
|
| 119 |
+
# Формируем выходные данные
|
| 120 |
+
usernames_output = "\n".join(usernames)
|
| 121 |
+
comments_output = "\n".join(comments)
|
| 122 |
+
likes_chronology_output = "\n".join(likes)
|
| 123 |
+
total_likes_sum = sum(map(int, likes))
|
| 124 |
+
|
| 125 |
+
# Подробная аналитика
|
| 126 |
+
analytics_summary = (
|
| 127 |
+
f"Content Type: {content_type}\n"
|
| 128 |
+
f"Link to Post: {link_to_post}\n\n"
|
| 129 |
+
f"ОСНОВНАЯ СТАТИСТИКА:\n"
|
| 130 |
+
f"- Всего комментариев: {total_comments}\n"
|
| 131 |
+
f"- Всего лайков на комментариях: {total_likes_sum}\n"
|
| 132 |
+
f"- Среднее количество лайков: {avg_likes:.1f}\n"
|
| 133 |
+
f"- Период активности: {earliest_week}-{latest_week} недель\n\n"
|
| 134 |
+
f"АНАЛИЗ КОНТЕНТА:\n"
|
| 135 |
+
f"- Средняя длина комментария: {avg_comment_length:.1f} символов\n"
|
| 136 |
+
f"- Всего эмодзи использовано: {total_emojis}\n"
|
| 137 |
+
f"- Тональность комментариев:\n"
|
| 138 |
+
f" * Позитивных: {sentiment_distribution['positive']}\n"
|
| 139 |
+
f" * Нейтральных: {sentiment_distribution['neutral']}\n"
|
| 140 |
+
f" * Негативных: {sentiment_distribution['negative']}\n\n"
|
| 141 |
+
f"АКТИВНОСТЬ ПОЛЬЗОВАТЕЛЕЙ:\n"
|
| 142 |
+
f"Самые активные комментаторы:\n"
|
| 143 |
+
+ "\n".join([f"- {user}: {count} комментариев" for user, count in most_active_users]) + "\n\n"
|
| 144 |
+
f"Самые упоминаемые пользователи:\n"
|
| 145 |
+
+ "\n".join([f"- {user}: {count} упоминаний" for user, count in most_mentioned if user]) + "\n\n"
|
| 146 |
+
f"ВОВЛЕЧЕННОСТЬ:\n"
|
| 147 |
+
f"- Процент комментариев с лайками: {(sum(1 for l in likes if int(l) > 0) / total_comments * 100):.1f}%\n"
|
| 148 |
+
f"- Процент комментариев с эмодзи: {(sum(1 for c in comments if count_emojis(c) > 0) / total_comments * 100):.1f}%\n"
|
| 149 |
+
)
|
| 150 |
+
|
| 151 |
+
return analytics_summary, usernames_output, comments_output, likes_chronology_output, str(total_likes_sum)
|
| 152 |
+
|
| 153 |
+
except Exception as e:
|
| 154 |
+
error_message = f"Произошла ошибка при обработке: {str(e)}\n{str(type(e))}"
|
| 155 |
+
return error_message, error_message, error_message, error_message, "0"
|
| 156 |
|
| 157 |
# Создаем интерфейс Gradio
|
| 158 |
iface = gr.Interface(
|
| 159 |
fn=analyze_post,
|
| 160 |
inputs=[
|
| 161 |
+
gr.Radio(
|
| 162 |
+
choices=["Photo", "Video"],
|
| 163 |
+
label="Content Type",
|
| 164 |
+
value="Photo"
|
| 165 |
+
),
|
| 166 |
+
gr.Textbox(
|
| 167 |
+
label="Link to Post",
|
| 168 |
+
placeholder="Введите ссылку на пост"
|
| 169 |
+
),
|
| 170 |
+
gr.Number(
|
| 171 |
+
label="Likes",
|
| 172 |
+
value=0
|
| 173 |
+
),
|
| 174 |
+
gr.Textbox(
|
| 175 |
+
label="Post Date",
|
| 176 |
+
placeholder="Введите дату публикации"
|
| 177 |
+
),
|
| 178 |
+
gr.Textbox(
|
| 179 |
+
label="Description",
|
| 180 |
+
placeholder="Введите описание поста",
|
| 181 |
+
lines=3
|
| 182 |
+
),
|
| 183 |
+
gr.Number(
|
| 184 |
+
label="Total Comment Count",
|
| 185 |
+
value=0
|
| 186 |
+
),
|
| 187 |
+
gr.Textbox(
|
| 188 |
+
label="All Comments",
|
| 189 |
+
placeholder="Вставьте комментарии",
|
| 190 |
+
lines=10
|
| 191 |
+
)
|
| 192 |
],
|
| 193 |
outputs=[
|
| 194 |
+
gr.Textbox(label="Analytics Summary", lines=20),
|
| 195 |
+
gr.Textbox(label="Usernames (Output 1)", lines=5),
|
| 196 |
+
gr.Textbox(label="Comments (Output 2)", lines=5),
|
| 197 |
+
gr.Textbox(label="Likes Chronology (Output 3)", lines=5),
|
| 198 |
+
gr.Textbox(label="Total Likes on Comments (Output 4)")
|
| 199 |
],
|
| 200 |
+
title="Instagram Comment Analyzer Pro",
|
| 201 |
+
description="Расширенный анализатор комментариев Instagram с детальной аналитикой"
|
| 202 |
)
|
| 203 |
|
| 204 |
+
if __name__ == "__main__":
|
| 205 |
+
iface.launch()
|