MS Chart Control中keyword的格式化

文章详细介绍了如何在饼图中显示销量数据时,使用格式化字符串给大数字销量加上千分位,同时仅显示整数部分,方便用户识别。

摘要生成于 C知道 ,由 DeepSeek-R1 满血版支持, 前往体验 >

前文中提到用keyword解决了饼图标签中同时显示数值和占比的问题,方法类似这样:

chart1.Series[0].Label = "#VALY[#PERCENT]";

比如,我现在的饼图用来展示公司08年各类产品的销量及占百分比,那么,其中#VALY指定显示的值就是销量,而#PERCENT显示的就是各类产品销量占总销量的百分比。

现在又出现一个问题,销量可能是个很大的数字,比如:123456789,这时,为了方便识别,就需要加上千分位。

怎么解决这个问题呢?原来,对于keyword,也支持格式化字符串,其规则和其它地方的用法类似。比如:

chart1.Series[0].Label = "#VALY{n0}[PERCENT]";

就可以给销量加上千分位了,同时只显示整数。

转自:http://blog.youkuaiyun.com/gaofeng2000/article/details/3985669

不过我按上面的写法却不行:
后来改成: this.ChartBusiness.Series[0].Label = "#VALX #PERCENT{P} #VALY";
OK了,可是千分位呢?

chart1.Series[0].Label = "#VALY[#PERCENT]";

比如,我现在的饼图用来展示公司08年各类产品的销量及占百分比,那么,其中#VALY指定显示的值就是销量,而#PERCENT显示的就是各类产品销量占总销量的百分比。

现在又出现一个问题,销量可能是个很大的数字,比如:123456789,这时,为了方便识别,就需要加上千分位。

怎么解决这个问题呢?原来,对于keyword,也支持格式化字符串,其规则和其它地方的用法类似。比如:

chart1.Series[0].Label = "#VALY{n0}[PERCENT]";

就可以给销量加上千分位了,同时只显示整数。

转自:http://blog.youkuaiyun.com/gaofeng2000/article/details/3985669

不过我按上面的写法却不行:后来改成: this.ChartBusiness.Series[0].Label = "#VALX #PERCENT{P} #VALY";OK了,可是千分位呢?
检查代码并优化: import sys import os import json import time import wave import numpy as np import pandas as pd import matplotlib.pyplot as plt import soundfile as sf # 新增依赖,用于音频片段保存 from matplotlib.backends.backend_qt5agg import FigureCanvasQTAgg as FigureCanvas from PyQt5.QtWidgets import (QApplication, QMainWindow, QWidget, QVBoxLayout, QHBoxLayout, QPushButton, QLabel, QLineEdit, QTextEdit, QFileDialog, QProgressBar, QGroupBox, QComboBox, QCheckBox) from PyQt5.QtCore import QThread, pyqtSignal from pydub import AudioSegment from transformers import pipeline, AutoTokenizer, AutoModelForSequenceClassification import whisper from pyannote.audio import Pipeline from docx import Document from docx.shared import Inches import librosa import tempfile from collections import defaultdict import re # 新增用于中文字符计数 class AnalysisThread(QThread): progress = pyqtSignal(int) message = pyqtSignal(str) analysis_complete = pyqtSignal(dict) error = pyqtSignal(str) def __init__(self, audio_files, keyword_file, whisper_model_path, pyannote_model_path, emotion_model_path): super().__init__() self.audio_files = audio_files self.keyword_file = keyword_file self.whisper_model_path = whisper_model_path self.pyannote_model_path = pyannote_model_path self.emotion_model_path = emotion_model_path self.running = True self.cached_models = {} def run(self): try: # 加载关键词 self.message.emit("正在加载关键词...") keywords = self.load_keywords() # 预加载模型 self.message.emit("正在预加载模型...") self.preload_models() results = [] total_files = len(self.audio_files) for idx, audio_file in enumerate(self.audio_files): if not self.running: self.message.emit("分析已停止") return self.message.emit(f"正在处理文件: {os.path.basename(audio_file)} ({idx+1}/{total_files})") file_result = self.analyze_file(audio_file, keywords) if file_result: results.append(file_result) self.progress.emit(int((idx + 1) / total_files * 100)) self.analysis_complete.emit({"results": results, "keywords": keywords}) self.message.emit("分析完成!") except Exception as e: import traceback self.error.emit(f"分析过程中发生错误: {str(e)}\n{traceback.format_exc()}") def preload_models(self): """预加载所有模型到缓存""" # 检查是否已加载模型 if hasattr(self, 'cached_models') and self.cached_models: return self.cached_models = {} # 加载语音识别模型 if 'whisper' not in self.cached_models: self.message.emit("正在加载语音识别模型...") self.cached_models['whisper'] = whisper.load_model(self.whisper_model_path) # 加载说话人分离模型 if 'pyannote' not in self.cached_models: self.message.emit("正在加载说话人分离模型...") self.cached_models['pyannote'] = Pipeline.from_pretrained(self.pyannote_model_path) # 加载情感分析模型 if 'emotion_classifier' not in self.cached_models: self.message.emit("正在加载情感分析模型...") tokenizer = AutoTokenizer.from_pretrained(self.emotion_model_path) model = AutoModelForSequenceClassification.from_pretrained(self.emotion_model_path) self.cached_models['emotion_classifier'] = pipeline( "text-classification", model=model, tokenizer=tokenizer, device=0 if torch.cuda.is_available() else -1 # 使用GPU如果可用 ) def analyze_file(self, audio_file, keywords): """分析单个音频文件""" try: # 确保音频为WAV格式 wav_file = self.convert_to_wav(audio_file) # 获取音频信息 duration, sample_rate, channels = self.get_audio_info(wav_file) # 说话人分离 diarization = self.cached_models['pyannote'](wav_file) # 识别客服和客户(使用改进的方法) agent_segments, customer_segments = self.identify_speakers(wav_file, diarization, keywords['opening']) # 语音识别(使用优化后的方法) whisper_model = self.cached_models['whisper'] agent_text = self.transcribe_audio(wav_file, agent_segments, whisper_model) customer_text = self.transcribe_audio(wav_file, customer_segments, whisper_model) # 情感分析 emotion_classifier = self.cached_models['emotion_classifier'] agent_emotion = self.analyze_emotion(agent_text, emotion_classifier) customer_emotion = self.analyze_emotion(customer_text, emotion_classifier) # 服务规范检查 opening_check = self.check_opening(agent_text, keywords['opening']) closing_check = self.check_closing(agent_text, keywords['closing']) forbidden_check = self.check_forbidden(agent_text, keywords['forbidden']) # 沟通技巧分析(使用改进的方法) speech_rate = self.analyze_speech_rate(wav_file, agent_segments) volume_analysis = self.analyze_volume(wav_file, agent_segments) # 问题解决率分析 resolution_rate = self.analyze_resolution(agent_text, customer_text, keywords['resolution']) # 构建结果 return { "file_name": os.path.basename(audio_file), "duration": duration, "agent_text": agent_text, "customer_text": customer_text, "opening_check": opening_check, "closing_check": closing_check, "forbidden_check": forbidden_check, "agent_emotion": agent_emotion, "customer_emotion": customer_emotion, "speech_rate": speech_rate, "volume_mean": volume_analysis['mean'], "volume_std": volume_analysis['std'], "resolution_rate": resolution_rate } except Exception as e: self.error.emit(f"处理文件 {os.path.basename(audio_file)} 时出错: {str(e)}") return None def load_keywords(self): """从Excel文件加载关键词""" try: df = pd.read_excel(self.keyword_file) keywords = { "opening": [str(k).strip() for k in df['opening'].dropna().tolist()], "closing": [str(k).strip() for k in df['closing'].dropna().tolist()], "forbidden": [str(k).strip() for k in df['forbidden'].dropna().tolist()], "resolution": [str(k).strip() for k in df['resolution'].dropna().tolist()] } return keywords except Exception as e: raise Exception(f"加载关键词文件失败: {str(e)}") def convert_to_wav(self, audio_file): """将音频文件转换为WAV格式(如果需要)""" try: if not audio_file.lower().endswith('.wav'): # 使用临时文件避免磁盘IO with tempfile.NamedTemporaryFile(suffix='.wav', delete=False) as tmpfile: output_file = tmpfile.name audio = AudioSegment.from_file(audio_file) audio.export(output_file, format='wav') return output_file return audio_file except Exception as e: raise Exception(f"音频转换失败: {str(e)}") def get_audio_info(self, wav_file): """获取音频文件信息""" try: with wave.open(wav_file, 'rb') as wf: frames = wf.getnframes() rate = wf.getframerate() channels = wf.getnchannels() duration = frames / float(rate) return duration, rate, channels except Exception as e: raise Exception(f"获取音频信息失败: {str(e)}") def identify_speakers(self, wav_file, diarization, opening_keywords): """改进的客服识别方法 - 检查前三个片段是否有开场白关键词""" speaker_segments = defaultdict(list) for segment, _, speaker in diarization.itertracks(yield_label=True): speaker_segments[speaker].append((segment.start, segment.end)) # 如果没有说话人 if not speaker_segments: return [], [] # 如果只有一个说话人 if len(speaker_segments) == 1: speaker = list(speaker_segments.keys())[0] return speaker_segments[speaker], [] # 检查每个说话人的前三个片段是否有开场白 speaker_scores = {} whisper_model = self.cached_models['whisper'] for speaker, segments in speaker_segments.items(): score = 0 # 取前三个片段(或所有片段如果少于3个) check_segments = segments[:3] for start, end in check_segments: # 转录片段 text = self.transcribe_audio_segment(wav_file, [(start, end)], whisper_model) # 检查开场白关键词 for keyword in opening_keywords: if keyword and keyword in text: score += 1 break speaker_scores[speaker] = score # 找到得分最高的说话人作为客服 agent_speaker = max(speaker_scores, key=speaker_scores.get) agent_segments = [] customer_segments = [] for speaker, segments in speaker_segments.items(): if speaker == agent_speaker: agent_segments = segments else: customer_segments.extend(segments) return agent_segments, customer_segments def transcribe_audio_segment(self, wav_file, segments, model): """转录单个音频片段 - 用于客服识别""" if not segments: return "" # 使用pydub加载音频 audio = AudioSegment.from_wav(wav_file) start, end = segments[0] # 转换为毫秒 start_ms = int(start * 1000) end_ms = int(end * 1000) segment_audio = audio[start_ms:end_ms] # 使用临时文件 with tempfile.NamedTemporaryFile(suffix='.wav') as tmpfile: segment_audio.export(tmpfile.name, format="wav") result = model.transcribe(tmpfile.name) return result['text'] def transcribe_audio(self, wav_file, segments, model): """优化后的转录方法 - 按片段转录""" if not segments: return "" # 使用pydub加载音频 audio = AudioSegment.from_wav(wav_file) full_text = "" # 只处理指定片段 for start, end in segments: # 转换为毫秒 start_ms = int(start * 1000) end_ms = int(end * 1000) segment_audio = audio[start_ms:end_ms] # 使用临时文件避免内存占用 with tempfile.NamedTemporaryFile(suffix='.wav') as tmpfile: segment_audio.export(tmpfile.name, format="wav") result = model.transcribe(tmpfile.name) full_text += result['text'] + " " return full_text.strip() def analyze_emotion(self, text, classifier): """分析文本情感""" if not text.strip(): return {"label": "中性", "score": 0.0} # 截断长文本以提高性能 if len(text) > 500: text = text[:500] result = classifier(text, truncation=True, max_length=512) return { "label": result[0]['label'], "score": result[0]['score'] } def check_opening(self, text, opening_keywords): """检查开场白""" return any(keyword in text for keyword in opening_keywords if keyword) def check_closing(self, text, closing_keywords): """检查结束语""" return any(keyword in text for keyword in closing_keywords if keyword) def check_forbidden(self, text, forbidden_keywords): """检查服务禁语""" return any(keyword in text for keyword in forbidden_keywords if keyword) def analyze_speech_rate(self, wav_file, segments): """改进的语速分析 - 基于实际识别文本""" if not segments: return 0 # 加载音频 y, sr = librosa.load(wav_file, sr=None) total_chars = 0 total_duration = 0 whisper_model = self.cached_models['whisper'] for start, end in segments: # 计算片段时长(秒) duration = end - start total_duration += duration # 转录片段 text = self.transcribe_audio_segment(wav_file, [(start, end)], whisper_model) # 计算中文字符数(去除标点和空格) chinese_chars = sum(1 for char in text if '\u4e00' <= char <= '\u9fff') total_chars += chinese_chars if total_duration == 0: return 0 # 语速 = 总字数 / 总时长(分钟) return total_chars / (total_duration / 60) def analyze_volume(self, wav_file, segments): """改进的音量分析 - 使用librosa计算RMS分贝值""" if not segments: return {"mean": -60, "std": 0} # 加载音频 y, sr = librosa.load(wav_file, sr=None) all_dB = [] for start, end in segments: start_sample = int(start * sr) end_sample = int(end * sr) segment_audio = y[start_sample:end_sample] # 计算RMS并转换为dB rms = librosa.feature.rms(y=segment_audio)[0] dB = librosa.amplitude_to_db(rms, ref=np.max) all_dB.extend(dB) if not all_dB: return {"mean": -60, "std": 0} return { "mean": float(np.mean(all_dB)), "std": float(np.std(all_dB)) } def analyze_resolution(self, agent_text, customer_text, resolution_keywords): """分析问题解决率""" return any(keyword in agent_text for keyword in resolution_keywords if keyword) def stop(self): """停止分析""" self.running = False class MainWindow(QMainWindow): def __init__(self): super().__init__() self.setWindowTitle("外呼电话录音包质检分析系统") self.setGeometry(100, 100, 1000, 700) # 初始化变量 self.audio_files = [] self.keyword_file = "" self.whisper_model_path = "./models/whisper-small" self.pyannote_model_path = "./models/pyannote-speaker-diarization" self.emotion_model_path = "./models/Erlangshen-Roberta-110M-Sentiment" self.output_dir = "./reports" # 创建主控件 central_widget = QWidget() self.setCentralWidget(central_widget) main_layout = QVBoxLayout(central_widget) # 文件选择区域 file_group = QGroupBox("文件选择") file_layout = QVBoxLayout(file_group) # 音频文件选择 audio_layout = QHBoxLayout() self.audio_label = QLabel("音频文件/文件夹:") audio_layout.addWidget(self.audio_label) self.audio_path_edit = QLineEdit() audio_layout.addWidget(self.audio_path_edit) self.audio_browse_btn = QPushButton("浏览...") self.audio_browse_btn.clicked.connect(self.browse_audio) audio_layout.addWidget(self.audio_browse_btn) file_layout.addLayout(audio_layout) # 关键词文件选择 keyword_layout = QHBoxLayout() self.keyword_label = QLabel("关键词文件:") keyword_layout.addWidget(self.keyword_label) self.keyword_path_edit = QLineEdit() keyword_layout.addWidget(self.keyword_path_edit) self.keyword_browse_btn = QPushButton("浏览...") self.keyword_browse_btn.clicked.connect(self.browse_keyword) keyword_layout.addWidget(self.keyword_browse_btn) file_layout.addLayout(keyword_layout) main_layout.addWidget(file_group) # 模型设置区域 model_group = QGroupBox("模型设置") model_layout = QVBoxLayout(model_group) # Whisper模型路径 whisper_layout = QHBoxLayout() whisper_layout.addWidget(QLabel("Whisper模型路径:")) self.whisper_edit = QLineEdit(self.whisper_model_path) whisper_layout.addWidget(self.whisper_edit) model_layout.addLayout(whisper_layout) # Pyannote模型路径 pyannote_layout = QHBoxLayout() pyannote_layout.addWidget(QLabel("Pyannote模型路径:")) self.pyannote_edit = QLineEdit(self.pyannote_model_path) pyannote_layout.addWidget(self.pyannote_edit) model_layout.addLayout(pyannote_layout) # 情感分析模型路径 emotion_layout = QHBoxLayout() emotion_layout.addWidget(QLabel("情感分析模型路径:")) self.emotion_edit = QLineEdit(self.emotion_model_path) emotion_layout.addWidget(self.emotion_edit) model_layout.addLayout(emotion_layout) # 输出目录 output_layout = QHBoxLayout() output_layout.addWidget(QLabel("输出目录:")) self.output_edit = QLineEdit(self.output_dir) output_layout.addWidget(self.output_edit) self.output_browse_btn = QPushButton("浏览...") self.output_browse_btn.clicked.connect(self.browse_output) output_layout.addWidget(self.output_browse_btn) model_layout.addLayout(output_layout) main_layout.addWidget(model_group) # 控制按钮区域 control_layout = QHBoxLayout() self.start_btn = QPushButton("开始分析") self.start_btn.clicked.connect(self.start_analysis) control_layout.addWidget(self.start_btn) self.stop_btn = QPushButton("停止分析") self.stop_btn.clicked.connect(self.stop_analysis) self.stop_btn.setEnabled(False) control_layout.addWidget(self.stop_btn) self.clear_btn = QPushButton("清空") self.clear_btn.clicked.connect(self.clear_all) control_layout.addWidget(self.clear_btn) main_layout.addLayout(control_layout) # 进度条 self.progress_bar = QProgressBar() self.progress_bar.setValue(0) main_layout.addWidget(self.progress_bar) # 日志输出区域 log_group = QGroupBox("分析日志") log_layout = QVBoxLayout(log_group) self.log_text = QTextEdit() self.log_text.setReadOnly(True) log_layout.addWidget(self.log_text) main_layout.addWidget(log_group) # 状态区域 status_layout = QHBoxLayout() self.status_label = QLabel("就绪") status_layout.addWidget(self.status_label) self.file_count_label = QLabel("已选择0个音频文件") status_layout.addWidget(self.file_count_label) main_layout.addLayout(status_layout) # 初始化分析线程 self.analysis_thread = None def browse_audio(self): """浏览音频文件或文件夹""" options = QFileDialog.Options() files, _ = QFileDialog.getOpenFileNames( self, "选择音频文件", "", "音频文件 (*.mp3 *.wav *.amr *.ogg *.flac);;所有文件 (*)", options=options ) if files: self.audio_files = files self.audio_path_edit.setText("; ".join(files)) self.file_count_label.setText(f"已选择{len(files)}个音频文件") self.log_text.append(f"已选择{len(files)}个音频文件") def browse_keyword(self): """浏览关键词文件""" options = QFileDialog.Options() file, _ = QFileDialog.getOpenFileName( self, "选择关键词文件", "", "Excel文件 (*.xlsx *.xls);;所有文件 (*)", options=options ) if file: self.keyword_file = file self.keyword_path_edit.setText(file) self.log_text.append(f"已选择关键词文件: {file}") def browse_output(self): """浏览输出目录""" options = QFileDialog.Options() directory = QFileDialog.getExistingDirectory( self, "选择输出目录", "", options=options ) if directory: self.output_dir = directory self.output_edit.setText(directory) self.log_text.append(f"输出目录设置为: {directory}") def start_analysis(self): """开始分析""" if not self.audio_files: self.log_text.append("错误: 请先选择音频文件") return if not self.keyword_file: self.log_text.append("错误: 请先选择关键词文件") return # 更新模型路径 self.whisper_model_path = self.whisper_edit.text() self.pyannote_model_path = self.pyannote_edit.text() self.emotion_model_path = self.emotion_edit.text() self.output_dir = self.output_edit.text() # 创建输出目录 os.makedirs(self.output_dir, exist_ok=True) self.log_text.append("开始分析...") self.start_btn.setEnabled(False) self.stop_btn.setEnabled(True) self.status_label.setText("分析中...") self.progress_bar.setValue(0) # 创建并启动分析线程 self.analysis_thread = AnalysisThread( self.audio_files, self.keyword_file, self.whisper_model_path, self.pyannote_model_path, self.emotion_model_path ) self.analysis_thread.progress.connect(self.progress_bar.setValue) self.analysis_thread.message.connect(self.log_text.append) self.analysis_thread.analysis_complete.connect(self.on_analysis_complete) self.analysis_thread.error.connect(self.on_analysis_error) self.analysis_thread.finished.connect(self.on_analysis_finished) self.analysis_thread.start() def stop_analysis(self): """停止分析""" if self.analysis_thread and self.analysis_thread.isRunning(): self.analysis_thread.stop() self.log_text.append("正在停止分析...") self.stop_btn.setEnabled(False) def clear_all(self): """清空所有内容""" self.audio_files = [] self.keyword_file = "" self.audio_path_edit.clear() self.keyword_path_edit.clear() self.log_text.clear() self.progress_bar.setValue(0) self.status_label.setText("就绪") self.file_count_label.setText("已选择0个音频文件") self.log_text.append("已清空所有内容") def on_analysis_complete(self, result): """分析完成处理""" try: self.log_text.append("正在生成报告...") if not result.get("results"): self.log_text.append("警告: 没有生成任何分析结果") return # 生成Excel报告 excel_path = os.path.join(self.output_dir, "质检分析报告.xlsx") self.generate_excel_report(result, excel_path) # 生成Word报告 word_path = os.path.join(self.output_dir, "质检分析报告.docx") self.generate_word_report(result, word_path) self.log_text.append(f"分析报告已保存至: {excel_path}") self.log_text.append(f"可视化报告已保存至: {word_path}") self.log_text.append("分析完成!") self.status_label.setText(f"分析完成!报告保存至: {self.output_dir}") except Exception as e: import traceback self.log_text.append(f"生成报告时出错: {str(e)}\n{traceback.format_exc()}") def on_analysis_error(self, message): """分析错误处理""" self.log_text.append(f"错误: {message}") self.status_label.setText("发生错误") def on_analysis_finished(self): """分析线程结束处理""" self.start_btn.setEnabled(True) self.stop_btn.setEnabled(False) def generate_excel_report(self, result, output_path): """生成Excel报告""" # 从结果中提取数据 data = [] for res in result['results']: data.append({ "文件名": res['file_name'], "音频时长(秒)": res['duration'], "开场白检查": "通过" if res['opening_check'] else "未通过", "结束语检查": "通过" if res['closing_check'] else "未通过", "服务禁语检查": "通过" if not res['forbidden_check'] else "未通过", "客服情感": res['agent_emotion']['label'], "客服情感得分": res['agent_emotion']['score'], "客户情感": res['customer_emotion']['label'], "客户情感得分": res['customer_emotion']['score'], "语速(字/分)": res['speech_rate'], "平均音量(dB)": res['volume_mean'], "音量标准差": res['volume_std'], "问题解决率": "是" if res['resolution_rate'] else "否" }) # 创建DataFrame并保存 df = pd.DataFrame(data) df.to_excel(output_path, index=False) # 添加汇总统计 try: with pd.ExcelWriter(output_path, engine='openpyxl', mode='a', if_sheet_exists='replace') as writer: summary_data = { "统计项": ["总文件数", "开场白通过率", "结束语通过率", "服务禁语通过率", "问题解决率"], "数值": [ len(result['results']), df['开场白检查'].value_counts().get('通过', 0) / len(df), df['结束语检查'].value_counts().get('通过', 0) / len(df), df['服务禁语检查'].value_counts().get('通过', 0) / len(df), df['问题解决率'].value_counts().get('是', 0) / len(df) ] } summary_df = pd.DataFrame(summary_data) summary_df.to_excel(writer, sheet_name='汇总统计', index=False) except Exception as e: self.log_text.append(f"添加汇总统计时出错: {str(e)}") def generate_word_report(self, result, output_path): """生成Word报告""" doc = Document() # 添加标题 doc.add_heading('外呼电话录音质检分析报告', 0) # 添加基本信息 doc.add_heading('分析概况', level=1) doc.add_paragraph(f"分析时间: {time.strftime('%Y-%m-%d %H:%M:%S')}") doc.add_paragraph(f"分析文件数量: {len(result['results'])}") doc.add_paragraph(f"关键词文件: {os.path.basename(self.keyword_file)}") # 添加汇总统计 doc.add_heading('汇总统计', level=1) # 创建汇总表格 table = doc.add_table(rows=5, cols=2) table.style = 'Table Grid' # 表头 hdr_cells = table.rows[0].cells hdr_cells[0].text = '统计项' hdr_cells[1].text = '数值' # 计算统计数据 df = pd.DataFrame(result['results']) pass_rates = { "开场白通过率": df['opening_check'].mean() if not df.empty else 0, "结束语通过率": df['closing_check'].mean() if not df.empty else 0, "服务禁语通过率": (1 - df['forbidden_check']).mean() if not df.empty else 0, "问题解决率": df['resolution_rate'].mean() if not df.empty else 0 } # 填充表格 rows = [ ("总文件数", len(result['results'])), ("开场白通过率", f"{pass_rates['开场白通过率']:.2%}"), ("结束语通过率", f"{pass_rates['结束语通过率']:.2%}"), ("服务禁语通过率", f"{pass_rates['服务禁语通过率']:.2%}"), ("问题解决率", f"{pass_rates['问题解决率']:.2%}") ] for i, row_data in enumerate(rows): if i < len(table.rows): row_cells = table.rows[i].cells row_cells[0].text = row_data[0] row_cells[1].text = str(row_data[1]) # 添加情感分析图表 if result['results']: doc.add_heading('情感分析', level=1) # 客服情感分布 agent_emotions = [res['agent_emotion']['label'] for res in result['results']] agent_emotion_counts = pd.Series(agent_emotions).value_counts() if not agent_emotion_counts.empty: fig, ax = plt.subplots(figsize=(6, 4)) agent_emotion_counts.plot.pie(autopct='%1.1f%%', ax=ax) ax.set_title('客服情感分布') plt.tight_layout() # 保存图表到临时文件 chart_path = os.path.join(self.output_dir, "agent_emotion_chart.png") plt.savefig(chart_path, dpi=100) plt.close() doc.add_picture(chart_path, width=Inches(4)) doc.add_paragraph('图1: 客服情感分布') # 客户情感分布 customer_emotions = [res['customer_emotion']['label'] for res in result['results']] customer_emotion_counts = pd.Series(customer_emotions).value_counts() if not customer_emotion_counts.empty: fig, ax = plt.subplots(figsize=(6, 4)) customer_emotion_counts.plot.pie(autopct='%1.1f%%', ax=ax) ax.set_title('客户情感分布') plt.tight_layout() chart_path = os.path.join(self.output_dir, "customer_emotion_chart.png") plt.savefig(chart_path, dpi=100) plt.close() doc.add_picture(chart_path, width=Inches(4)) doc.add_paragraph('图2: 客户情感分布') # 添加详细分析结果 doc.add_heading('详细分析结果', level=1) # 创建详细表格 table = doc.add_table(rows=1, cols=6) table.style = 'Table Grid' # 表头 hdr_cells = table.rows[0].cells headers = ['文件名', '开场白', '结束语', '禁语', '客服情感', '问题解决'] for i, header in enumerate(headers): hdr_cells[i].text = header # 填充数据 for res in result['results']: row_cells = table.add_row().cells row_cells[0].text = res['file_name'] row_cells[1].text = "✓" if res['opening_check'] else "✗" row_cells[2].text = "✓" if res['closing_check'] else "✗" row_cells[3].text = "✗" if res['forbidden_check'] else "✓" row_cells[4].text = res['agent_emotion']['label'] row_cells[5].text = "✓" if res['resolution_rate'] else "✗" # 保存文档 doc.save(output_path) if __name__ == "__main__": # 检查是否安装了torch try: import torch except ImportError: print("警告: PyTorch 未安装,情感分析可能无法使用GPU加速") app = QApplication(sys.argv) window = MainWindow()
07-15
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值