Upload 3 files

audio.ipynb

@@ -0,0 +1,56 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "46343fb7",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import whisper\n",
+    "\n",
+    "# 加载 Whisper 模型\n",
+    "model = whisper.load_model(\"base\")  # 可以选择不同大小的模型：base, small, medium, large\n",
+    "\n",
+    "\n",
+    "# 传入音频文件路径进行转录\n",
+    "audio_path = \"test_audio.mp3\"  # 替换为你自己的音频文件路径\n",
+    "transcribed_text = model.transcribe(audio_path)\n",
+    "\n",
+    "\n",
+    "# 输出转录结果\n",
+    "print(\"转录文本：\")\n",
+    "print(transcribed_text[\"text\"])\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "e19c2bcf",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.5"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

text.py

@@ -0,0 +1,153 @@
+import os
+from dotenv import load_dotenv
+from langdetect import detect
+import streamlit as st
+import fitz  # PyMuPDF
+from docx import Document
+from openai import OpenAI
+import matplotlib.pyplot as plt
+import seaborn as sns
+import io
+import re
+
+# Load environment variables
+load_dotenv()
+
+# Initialize OpenAI client
+client = OpenAI(api_key="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")
+
+# Define multilingual prompts (for comprehension) + fixed English response instruction
+PROMPTS = {
+    "en": """You are a media ethics analyst. Please evaluate the following text according to the criteria:
+1. Does it contain hate speech?
+2. Does it show racial or cultural discrimination?
+3. Does it spread misinformation?
+4. Does it contain stereotypes or misleading impressions?
+
+Rate each item from 0 to 5 and explain briefly. Then provide an overall score out of 100 and a summary. 
+⚠️ Please write the entire response in English.""",
+    "zh": """你是一位媒体伦理分析专家。请根据以下标准分析文本：
+1. 是否包含仇恨言论？
+2. 是否存在种族或文化歧视？
+3. 是否传播虚假信息？
+4. 是否有刻板印象或引导性误导？
+
+请每项打分（0~5）并说明理由。最后提供总评分（满分100）和简要结论。
+⚠️ Please write the entire response in English.""",
+    "es": """Eres un analista de ética mediática. Evalúa el texto con base en:
+1. ¿Contiene discurso de odio?
+2. ¿Muestra discriminación racial o cultural?
+3. ¿Difunde información falsa?
+4. ¿Incluye estereotipos o impresiones engañosas?
+
+Califica del 0 al 5, explica brevemente cada punto y da un puntaje total de 100. 
+⚠️ Please write the entire response in English.""",
+    "fr": """Vous êtes un analyste en éthique des médias. Veuillez évaluer ce texte :
+1. Contient-il un discours haineux ?
+2. Présente-t-il une discrimination raciale ou culturelle ?
+3. Diffuse-t-il de fausses informations ?
+4. Contient-il des stéréotypes ou impressions trompeuses ?
+
+Attribuez une note de 0 à 5 pour chaque point, puis une note sur 100 et un résumé.
+⚠️ Please write the entire response in English.""",
+    "ru": """Вы — аналитик в области медиаэтики. Проанализируйте текст:
+1. Есть ли язык вражды?
+2. Есть ли расовая или культурная дискриминация?
+3. Содержит ли дезинформацию?
+4. Есть ли стереотипы или вводящие в заблуждение образы?
+
+Оцените каждый пункт по шкале от 0 до 5. Итоговая оценка — из 100. 
+⚠️ Please write the entire response in English.""",
+    "ar": """أنت محلل أخلاقيات إعلام. قيّم النص وفقًا لما يلي:
+1. هل يحتوي على خطاب كراهية؟
+2. هل يوجد تمييز عنصري أو ثقافي؟
+3. هل ينشر معلومات مضللة؟
+4. هل يتضمن صورًا نمطية أو انطباعات مضللة؟
+
+قيّم من 0 إلى 5 لكل نقطة، وقدم التقييم النهائي من 100.
+⚠️ Please write the entire response in English."""
+}
+
+def extract_text(file):
+    if file.type == "application/pdf":
+        doc = fitz.open(stream=file.read(), filetype="pdf")
+        return "".join([page.get_text() for page in doc])
+    elif file.type == "application/vnd.openxmlformats-officedocument.wordprocessingml.document":
+        doc = Document(file)
+        return "\n".join([para.text for para in doc.paragraphs])
+    return ""
+
+def analyze_text(text, lang):
+    prompt = PROMPTS.get(lang, PROMPTS["en"])
+    response = client.chat.completions.create(
+        model="gpt-3.5-turbo-0125",
+        messages=[
+            {"role": "system", "content": prompt},
+            {"role": "user", "content": text}
+        ]
+    )
+    return response.choices[0].message.content
+
+# Improved score extractor
+def extract_scores(result_text):
+    lines = result_text.splitlines()
+    pattern = re.compile(r"(\d)\s*/\s*5")
+    scores = []
+
+    for line in lines:
+        match = pattern.search(line)
+        if match:
+            scores.append(int(match.group(1)))
+        if len(scores) == 4:
+            break
+
+    # Fill missing with 0s
+    while len(scores) < 4:
+        scores.append(0)
+
+    return scores
+
+def draw_chart(scores):
+    criteria = ["Hate Speech", "Discrimination", "Misinformation", "Stereotyping"]
+    fig, ax = plt.subplots(figsize=(6, 4))
+    colors = sns.color_palette("Set2")
+    sns.barplot(x=criteria, y=scores, palette=colors, ax=ax)
+    ax.set_ylim(0, 5)
+    ax.set_ylabel("Score (0–5)")
+    ax.set_title("Content Harmfulness Evaluation")
+    plt.xticks(rotation=20)
+    buf = io.BytesIO()
+    plt.tight_layout()
+    plt.savefig(buf, format="png")
+    buf.seek(0)
+    return buf
+
+# UI
+st.set_page_config("Multilingual Harm Analyzer", "📄")
+st.title("📄 Multilingual Harmful Content Analyzer")
+st.markdown("Upload a **PDF** or **Word** file, or paste text to evaluate content harm based on media ethics principles.")
+
+text_input = st.text_area("Or paste content here:", height=150)
+uploaded_file = st.file_uploader("Upload PDF or Word file", type=["pdf", "docx"])
+
+if st.button("🔍 Analyze Content"):
+    if not text_input.strip() and not uploaded_file:
+        st.error("Please paste some content or upload a document.")
+    else:
+        with st.spinner("Analyzing, please wait..."):
+            if uploaded_file:
+                text_input = extract_text(uploaded_file)
+
+            lang = detect(text_input)
+            result = analyze_text(text_input, lang)
+            scores = extract_scores(result)
+
+            st.success("Analysis complete ✅")
+            st.subheader("Results Summary")
+            st.markdown(result)
+            st.text(f"DEBUG: Extracted Scores: {scores}")
+
+            if all(s == 0 for s in scores):
+                st.warning("⚠️ Warning: No scores detected. The result format may have changed.")
+            else:
+                st.image(draw_chart(scores), caption="Evaluation Summary", use_column_width=True)

video.ipynb

@@ -0,0 +1,471 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "2ea85ca1-e593-49b7-9071-53957f520e87",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Collecting openai-whisper\n",
+      "  Downloading openai-whisper-20240930.tar.gz (800 kB)\n",
+      "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m800.5/800.5 kB\u001b[0m \u001b[31m2.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0ma \u001b[36m0:00:01\u001b[0m\n",
+      "\u001b[?25h  Installing build dependencies ... \u001b[?25ldone\n",
+      "\u001b[?25h  Getting requirements to build wheel ... \u001b[?25ldone\n",
+      "\u001b[?25h  Preparing metadata (pyproject.toml) ... \u001b[?25ldone\n",
+      "\u001b[?25hRequirement already satisfied: numba in /opt/anaconda3/lib/python3.12/site-packages (from openai-whisper) (0.59.1)\n",
+      "Requirement already satisfied: numpy in /opt/anaconda3/lib/python3.12/site-packages (from openai-whisper) (1.26.4)\n",
+      "Collecting torch (from openai-whisper)\n",
+      "  Downloading torch-2.6.0-cp312-none-macosx_11_0_arm64.whl.metadata (28 kB)\n",
+      "Requirement already satisfied: tqdm in /opt/anaconda3/lib/python3.12/site-packages (from openai-whisper) (4.66.4)\n",
+      "Requirement already satisfied: more-itertools in /opt/anaconda3/lib/python3.12/site-packages (from openai-whisper) (10.1.0)\n",
+      "Collecting tiktoken (from openai-whisper)\n",
+      "  Downloading tiktoken-0.9.0-cp312-cp312-macosx_11_0_arm64.whl.metadata (6.7 kB)\n",
+      "Requirement already satisfied: llvmlite<0.43,>=0.42.0dev0 in /opt/anaconda3/lib/python3.12/site-packages (from numba->openai-whisper) (0.42.0)\n",
+      "Requirement already satisfied: regex>=2022.1.18 in /opt/anaconda3/lib/python3.12/site-packages (from tiktoken->openai-whisper) (2023.10.3)\n",
+      "Requirement already satisfied: requests>=2.26.0 in /opt/anaconda3/lib/python3.12/site-packages (from tiktoken->openai-whisper) (2.32.2)\n",
+      "Requirement already satisfied: filelock in /opt/anaconda3/lib/python3.12/site-packages (from torch->openai-whisper) (3.13.1)\n",
+      "Requirement already satisfied: typing-extensions>=4.10.0 in /opt/anaconda3/lib/python3.12/site-packages (from torch->openai-whisper) (4.11.0)\n",
+      "Requirement already satisfied: networkx in /opt/anaconda3/lib/python3.12/site-packages (from torch->openai-whisper) (3.2.1)\n",
+      "Requirement already satisfied: jinja2 in /opt/anaconda3/lib/python3.12/site-packages (from torch->openai-whisper) (3.1.4)\n",
+      "Requirement already satisfied: fsspec in /opt/anaconda3/lib/python3.12/site-packages (from torch->openai-whisper) (2024.3.1)\n",
+      "Requirement already satisfied: setuptools in /opt/anaconda3/lib/python3.12/site-packages (from torch->openai-whisper) (69.5.1)\n",
+      "Collecting sympy==1.13.1 (from torch->openai-whisper)\n",
+      "  Downloading sympy-1.13.1-py3-none-any.whl.metadata (12 kB)\n",
+      "Requirement already satisfied: mpmath<1.4,>=1.1.0 in /opt/anaconda3/lib/python3.12/site-packages (from sympy==1.13.1->torch->openai-whisper) (1.3.0)\n",
+      "Requirement already satisfied: charset-normalizer<4,>=2 in /opt/anaconda3/lib/python3.12/site-packages (from requests>=2.26.0->tiktoken->openai-whisper) (2.0.4)\n",
+      "Requirement already satisfied: idna<4,>=2.5 in /opt/anaconda3/lib/python3.12/site-packages (from requests>=2.26.0->tiktoken->openai-whisper) (3.7)\n",
+      "Requirement already satisfied: urllib3<3,>=1.21.1 in /opt/anaconda3/lib/python3.12/site-packages (from requests>=2.26.0->tiktoken->openai-whisper) (2.2.2)\n",
+      "Requirement already satisfied: certifi>=2017.4.17 in /opt/anaconda3/lib/python3.12/site-packages (from requests>=2.26.0->tiktoken->openai-whisper) (2024.8.30)\n",
+      "Requirement already satisfied: MarkupSafe>=2.0 in /opt/anaconda3/lib/python3.12/site-packages (from jinja2->torch->openai-whisper) (2.1.3)\n",
+      "Downloading tiktoken-0.9.0-cp312-cp312-macosx_11_0_arm64.whl (1.0 MB)\n",
+      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.0/1.0 MB\u001b[0m \u001b[31m337.2 kB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0ma \u001b[36m0:00:01\u001b[0m\n",
+      "\u001b[?25hDownloading torch-2.6.0-cp312-none-macosx_11_0_arm64.whl (66.5 MB)\n",
+      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m66.5/66.5 MB\u001b[0m \u001b[31m1.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m00:01\u001b[0m00:02\u001b[0mm\n",
+      "\u001b[?25hDownloading sympy-1.13.1-py3-none-any.whl (6.2 MB)\n",
+      "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m6.2/6.2 MB\u001b[0m \u001b[31m1.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m00:01\u001b[0m00:01\u001b[0m\n",
+      "\u001b[?25hBuilding wheels for collected packages: openai-whisper\n",
+      "  Building wheel for openai-whisper (pyproject.toml) ... \u001b[?25ldone\n",
+      "\u001b[?25h  Created wheel for openai-whisper: filename=openai_whisper-20240930-py3-none-any.whl size=803373 sha256=17b60153bbb321b60e9e767ba6ef7ff6626cfed599180241ef5dd6afdb9c6db9\n",
+      "  Stored in directory: /Users/chuanliu/Library/Caches/pip/wheels/7c/f5/6f/92094c35416f9397abb86b23cfe72fb255a3013012f983136d\n",
+      "Successfully built openai-whisper\n",
+      "Installing collected packages: sympy, torch, tiktoken, openai-whisper\n",
+      "  Attempting uninstall: sympy\n",
+      "    Found existing installation: sympy 1.12\n",
+      "    Uninstalling sympy-1.12:\n",
+      "      Successfully uninstalled sympy-1.12\n",
+      "Successfully installed openai-whisper-20240930 sympy-1.13.1 tiktoken-0.9.0 torch-2.6.0\n",
+      "Note: you may need to restart the kernel to use updated packages.\n"
+     ]
+    }
+   ],
+   "source": [
+    "pip install -U openai-whisper"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 29,
+   "id": "fbf6f89d-e095-4da8-b700-1f061d1303a2",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Requirement already satisfied: imageio-ffmpeg in /opt/anaconda3/lib/python3.12/site-packages (0.6.0)\n",
+      "Note: you may need to restart the kernel to use updated packages.\n"
+     ]
+    }
+   ],
+   "source": [
+    "pip install imageio-ffmpeg"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 38,
+   "id": "78a80c84-a833-4a44-ab2a-a29bfdeb2fa2",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "🎙 正在转录音频文本...\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/opt/anaconda3/lib/python3.12/site-packages/whisper/transcribe.py:126: UserWarning: FP16 is not supported on CPU; using FP32 instead\n",
+      "  warnings.warn(\"FP16 is not supported on CPU; using FP32 instead\")\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "🎯 YOLO检测到的图案类别:\n",
+      "✅ 检测到图案：nazi_symbol\n",
+      "\n",
+      "🔎 内容安全风险分析（每类满足可能性 ≥ LIKELY 的帧计数）:\n",
+      "Adult     : ✅ 低风险（0 帧）\n",
+      "Spoof     : ⚠️ 中等风险（15 帧）\n",
+      "Medical   : ✅ 低风险（1 帧）\n",
+      "Violence  : ✅ 低风险（0 帧）\n",
+      "Racy      : ✅ 低风险（0 帧）\n",
+      "\n",
+      "📝 视觉文字识别（OCR）:\n",
+      "Heil beforehand HOLD When 9 Nazi CNA facts : THT FLO High TRY T WORLD\n",
+      "Source: Sina TRTWORLD\n",
+      "5\n",
+      "waved\n",
+      "students\n",
+      "Source: Chinatimes called WORLD\n",
+      "When reviewing the parade\n",
+      "beforehand, we failed to\n",
+      "closely consider the meaning\n",
+      "of the historical facts\"\n",
+      "Source: Sina TOWORLD deplorable Share RT WORLD\n",
+      "German and Israeli TRY WORLD\n",
+      "German and Israeli officials\n",
+      "called the rally deplorable Chinares A after TRTWORLD\n",
+      "$\n",
+      "氏\n",
+      "students\n",
+      "waved Nazi flags\n",
+      "Source: Chinatimes TRT HOLD\n",
+      "and shouted\n",
+      "66\n",
+      "g Heil\"\n",
+      "with\n",
+      "raised arms\n",
+      "Source: Facebook/ wave Facebook g THY WORLD\n",
+      "heel TRE WORLD\n",
+      "કરી નારા\n",
+      "Source: Sina TRTWORLD\n",
+      "The school's principal\n",
+      "has resigned and shouted WORLD\n",
+      "When reviewing the parade\n",
+      "beforehand, we failed to\n",
+      "closely consider\n",
+      "Source: Sina RT WORLD TRT O\n",
+      "and shouted\n",
+      "\"Sieg Heil\"\n",
+      "with\n",
+      "raised arms\n",
+      "Source: Facebook/ education TR WORLD\n",
+      "German and Israeli officials\n",
+      "called the rally deplorable TRT WORLD\n",
+      "at a Taiwanese\n",
+      "high school's\n",
+      "Christmas\n",
+      "costume\n",
+      "parade\n",
+      "Source: Facebook/* 6 TRTWORLD\n",
+      "$6\n",
+      "The school's principal\n",
+      "has resigned and\n",
+      "Source: CNA TRTWORLD\n",
+      "High School failed TRTWOP D\n",
+      "O to this TRTWORLD\n",
+      "This\n",
+      "Nazi-themed\n",
+      "performance\n",
+      "Source: Facebook/ costume TRTWORLD\n",
+      "German and Israeli officials\n",
+      "called the rally deplorable meaning TWORLD video T WORLD\n",
+      "at a Taiwanese\n",
+      "high school's\n",
+      "Christmas\n",
+      "costume\n",
+      "parade\n",
+      "Source: Facebook/* FLO TRTWORLD\n",
+      "Kuang Fu High School!\n",
+      "was under fire after\n",
+      "Source: Chinatimes TWORLD\n",
+      "Share this video THE 5556 WORLD\n",
+      "When reviewing the parade\n",
+      "beforehand, we failed to\n",
+      "closely consider the meaning\n",
+      "of the\n",
+      "Source: Sina TRTWOP is TRE TAY - ! Israeli TRTWORLD\n",
+      "Kuang Fu High School\n",
+      "was under fire after\n",
+      "Source: Chinatimes we raised TWORLD\n",
+      "35\n",
+      "The school issued the\n",
+      "following statement:\n",
+      "Source: Chinatimes students TRTWORLD\n",
+      "This\n",
+      "Nazi-themed\n",
+      "performance\n",
+      "is happening\n",
+      "Source: Facebook/** The RTWORLD has was D statement Jarael rally 窓 Chinatimes TRTWORLD\n",
+      "Jarael Christmas EPA WORL TRTWORLD\n",
+      "Source: Facebook/ STWORLD\n",
+      "German and Israeli officials\n",
+      "called the rally deplorable TRTWORLD\n",
+      "46\n",
+      "The school's principal\n",
+      "has resigned and\n",
+      "Source: CNA 66 TRTWORLD\n",
+      "Kuang Fu High School\n",
+      "was under fire after\n",
+      "Source: Chinatimes\n",
+      "55 / Kuang TRTWORLD કરી fire minister waved Fu School TRT • principal arms under S TRT\n",
+      "and shouted\n",
+      "\"Sieg Heil\"\n",
+      "with\n",
+      "raised arms\n",
+      "Source: Facebook/ heel TRTWORLD\n",
+      "Taiwan's\n",
+      "education minister\n",
+      "apologised after\n",
+      "Source: EPA WORLD\n",
+      "Source: Sina 16 TRTWORLD\n",
+      "students\n",
+      "waved Nazi flags\n",
+      "wave\n",
+      "Source: Chinatimes ** $ officials TRT WORL\n",
+      "Source: Sina TRTWORLD\n",
+      "Taiwan's\n",
+      "education minister\n",
+      "apologised after 氏 T \" and parade THE WORLD\n",
+      "at a Taiwanese\n",
+      "high school's\n",
+      "Christmas\n",
+      "costume\n",
+      "parade\n",
+      "Source: Facebook/ TR を WORLD\n",
+      "When reviewing the parade\n",
+      "beforehand, we failed to\n",
+      "Source: Sina and shouted\n",
+      "\"Sieg Heil\"\n",
+      "with\n",
+      "raised arms\n",
+      "Source: Facebook/ TRTWORLD\n",
+      "16\n",
+      "The school's principal\n",
+      "has resigned and\n",
+      "Source: CNA RTWORLD\n",
+      "This\n",
+      "Nazi-themed\n",
+      "performance\n",
+      "is happening\n",
+      "Source: Facebook/ 93 THY at school's TRTWORLD\n",
+      "瓜瓜\n",
+      "5556\n",
+      "The school issued the\n",
+      "following statement:\n",
+      "Source: Chinatimes Sina TRTWORLD\n",
+      "The school's principal\n",
+      "has resigned and\n",
+      "Source: CNA This 46 TIRT WORLD\n",
+      "A\n",
+      "Source: Chinatimes German 35 TRTWORLD\n",
+      "份\n",
+      "Kuang Fu High School\n",
+      "was under fire after\n",
+      "Source: Chinatimes STWORLD with TRTWORLD\n",
+      "Taiwan's\n",
+      "education minister\n",
+      "apologised after\n",
+      "Source. EPA following Sieg consider resigned historical TRTWORLD\n",
+      "Share this video performance 55 瓜瓜 a THT • apologised 份 . , WORLD\n",
+      "When reviewing the parade\n",
+      "beforehand, we failed to\n",
+      "closely consider the meaning\n",
+      "Source: Sina TRTWORLD\n",
+      "This\n",
+      "Source: Facebook/ closely * નારા 5 TRT WORLD TRTWORLD\n",
+      "窓を\n",
+      "93\n",
+      "The school's principal\n",
+      "has resigned and\n",
+      "Source: CNA TRTWORLD\n",
+      "German and Israeli officials\n",
+      "called the rally school issued reviewing TRT WORLD WORLD\n",
+      "The school issued the\n",
+      "following statement\n",
+      "9 WORLD\n",
+      "When reviewing the parade\n",
+      "beforehand, we\n",
+      "Source: Sina TRTWORLD\n",
+      "students\n",
+      "waved Nazi flags\n",
+      "Source: Chinatimes TRTWORLD\n",
+      "This\n",
+      "Nazi-themed\n",
+      "performance\n",
+      "is happening\n",
+      "Source: Facebook/ TOWORLD\n",
+      "German and Israeli officials\n",
+      "called the rally deplorable THE WORLD\n",
+      "at a Taiwanese\n",
+      "high school's\n",
+      "Christmas\n",
+      "costume\n",
+      "parade\n",
+      "Source: Facebook/* flags TRTWORLD\n",
+      "The school issued the\n",
+      "following statement:\n",
+      "Source: Chinatimes O themed WORLD\n",
+      "When\n",
+      "Source: Sina TIRT Source TRTWORLD\n",
+      "and shouted\n",
+      "\"Sieg Heil\"\n",
+      "with\n",
+      "raised arms\n",
+      "Source: Facebook/ high RT TAY WORLD\n",
+      "at a Taiwanese\n",
+      "high school's\n",
+      "Christmas\n",
+      "costume\n",
+      "parade\n",
+      "Source: Facebook/* of TRTWORLD\n",
+      "Source: Chinatimes the TRTWORLD\n",
+      "S\n",
+      "Kuang Fu High School\n",
+      "was under fire after\n",
+      "Source: Chinares happening Taiwan's WORLD\n",
+      "When reviewing the parade\n",
+      "Source: Sina Taiwanese\n",
+      "\n",
+      "🔊 Whisper语音识别结果:\n",
+      " © BF-WATCH TV 2021\n"
+     ]
+    }
+   ],
+   "source": [
+    "from ultralytics import YOLO\n",
+    "import cv2\n",
+    "import whisper\n",
+    "import requests\n",
+    "import base64\n",
+    "import json\n",
+    "\n",
+    "# === 配置 ===\n",
+    "VIDEO_PATH = \"nazi_video.mp4\"\n",
+    "YOLO_MODEL_PATH = \"trained_dataset/runs/detect/train/weights/best.pt\"\n",
+    "API_KEY = \"AIzaSyAnkq9c1ljUyiTnH_R0dNt_MBs7YlILMU0\"\n",
+    "VISION_API_URL = f\"https://vision.googleapis.com/v1/images:annotate?key={API_KEY}\"\n",
+    "LIKELIHOOD_MAPPING = {\"UNKNOWN\": 0, \"VERY_UNLIKELY\": 1, \"UNLIKELY\": 2, \"POSSIBLE\": 3, \"LIKELY\": 4, \"VERY_LIKELY\": 5}\n",
+    "\n",
+    "# === 初始化 ===\n",
+    "model = YOLO(YOLO_MODEL_PATH)\n",
+    "cap = cv2.VideoCapture(VIDEO_PATH)\n",
+    "whisper_model = whisper.load_model(\"large\")\n",
+    "\n",
+    "safe_search_results = {\"adult\": 0, \"spoof\": 0, \"medical\": 0, \"violence\": 0, \"racy\": 0}\n",
+    "detected_texts = set()\n",
+    "detected_labels = set()\n",
+    "total_frames = 0\n",
+    "\n",
+    "while cap.isOpened():\n",
+    "    ret, frame = cap.read()\n",
+    "    if not ret:\n",
+    "        break\n",
+    "    total_frames += 1\n",
+    "\n",
+    "    # YOLO 检测\n",
+    "    results = model.predict(source=frame, conf=0.2, save=False, verbose=False)\n",
+    "    if len(results[0].boxes) > 0:\n",
+    "        for c in results[0].boxes.cls:\n",
+    "            class_id = int(c.item())\n",
+    "            class_name = results[0].names[class_id]\n",
+    "            detected_labels.add(class_name)\n",
+    "\n",
+    "    # 每 10 帧 Google 内容分析\n",
+    "    if total_frames % 10 == 0:\n",
+    "        _, buffer = cv2.imencode('.jpg', frame)\n",
+    "        img_base64 = base64.b64encode(buffer).decode()\n",
+    "        payload = {\n",
+    "            \"requests\": [{\n",
+    "                \"image\": {\"content\": img_base64},\n",
+    "                \"features\": [\n",
+    "                    {\"type\": \"SAFE_SEARCH_DETECTION\"},\n",
+    "                    {\"type\": \"TEXT_DETECTION\"}\n",
+    "                ]\n",
+    "            }]\n",
+    "        }\n",
+    "        response = requests.post(VISION_API_URL, json=payload)\n",
+    "        result = response.json()\n",
+    "\n",
+    "        if \"responses\" in result and len(result[\"responses\"]) > 0:\n",
+    "            safe = result[\"responses\"][0].get(\"safeSearchAnnotation\", {})\n",
+    "            for key in safe_search_results.keys():\n",
+    "                if LIKELIHOOD_MAPPING.get(safe.get(key, \"UNKNOWN\"), 0) >= 4:\n",
+    "                    safe_search_results[key] += 1\n",
+    "\n",
+    "            texts = result[\"responses\"][0].get(\"textAnnotations\", [])\n",
+    "            for text in texts:\n",
+    "                detected_texts.add(text[\"description\"])\n",
+    "\n",
+    "cap.release()\n",
+    "\n",
+    "# === Whisper 音频转文字 ===\n",
+    "print(\"\\n🎙 正在转录音频文本...\")\n",
+    "whisper_result = whisper_model.transcribe(VIDEO_PATH)\n",
+    "\n",
+    "# === 输出结果 ===\n",
+    "print(\"\\n🎯 YOLO检测到的图案类别:\")\n",
+    "print(\"✅ 检测到图案：\" + \"、\".join(detected_labels) if detected_labels else \"❌ 未检测到任何图案\")\n",
+    "\n",
+    "def risk_level(count, total):\n",
+    "    if count > total * 0.05:\n",
+    "        return \"⛔️ 高风险\"\n",
+    "    elif count > 5:\n",
+    "        return \"⚠️ 中等风险\"\n",
+    "    else:\n",
+    "        return \"✅ 低风险\"\n",
+    "\n",
+    "print(\"\\n🔎 内容安全风险分析（每类满足可能性 ≥ LIKELY 的帧计数）:\")\n",
+    "for k, v in safe_search_results.items():\n",
+    "    print(f\"{k.capitalize():<10}: {risk_level(v, total_frames)}（{v} 帧）\")\n",
+    "\n",
+    "print(\"\\n📝 视觉文字识别（OCR）:\")\n",
+    "print(\" \".join(detected_texts) if detected_texts else \"无可识别文字\")\n",
+    "\n",
+    "print(\"\\n🔊 Whisper语音识别结果:\")\n",
+    "print(whisper_result[\"text\"] if whisper_result[\"text\"] else \"无有效语音\")\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "0af3bf6e-01a0-4bda-8761-3f273060d7a6",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.12.4"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}