Update pitch_animation.py

pitch_animation.py

@@ -3,7 +3,7 @@ import re
 import sys
 import numpy as np
 import matplotlib
-# 强制使用 Agg 后端，确保云端环境渲染稳定
+# 强制使用 Agg 后端
 matplotlib.use('Agg')
 import matplotlib.pyplot as plt
 import matplotlib.animation as animation
@@ -19,7 +19,6 @@ class PitchAnimationGenerator:
         self.output_dir = output_video_dir
         self.fps = 15
         self.dpi = 100
-        # 获取系统字体
         self.zh_font = self._get_system_font()
         
         # 音名映射表 (MIDI 0-127)
@@ -28,7 +27,6 @@ class PitchAnimationGenerator:
         
         os.makedirs(self.output_dir, exist_ok=True)
         
-        # 数据加载流水线
         self.summary_segments = self._parse_summary_report()
         self.all_pitch_info = self._parse_pitch_srt()
         self.pitch_segments = self._match_segment_and_pitch()
@@ -38,7 +36,6 @@ class PitchAnimationGenerator:
         linux_path = '/usr/share/fonts/truetype/wqy/wqy-microhei.ttc'
         mac_path = '/System/Library/Fonts/PingFang.ttc'
         win_path = 'C:/Windows/Fonts/msyh.ttc'
-        
         if os.path.exists(linux_path): return FontProperties(fname=linux_path, size=10)
         if os.path.exists(mac_path): return FontProperties(fname=mac_path, size=10)
         if os.path.exists(win_path): return FontProperties(fname=win_path, size=10)
@@ -48,8 +45,7 @@ class PitchAnimationGenerator:
         try:
             h, m, s = t.replace(',', '.').split(':')
             return int(h)*3600 + int(m)*60 + float(s)
-        except:
-            return 0.0
+        except: return 0.0
 
     def _note_to_midi(self, note_str):
         if not note_str: return 60
@@ -69,13 +65,11 @@ class PitchAnimationGenerator:
                     if match:
                         t_range = match.group(2)
                         times = t_range.split(' --> ')
-                        start, end = self._time_to_sec(times[0]), self._time_to_sec(times[1])
                         segments.append({
                             'seq_num': int(match.group(1)), 
                             'time_range': t_range, 
-                            'start_sec': start, 
-                            'end_sec': end, 
-                            'duration': end - start
+                            'start_sec': self._time_to_sec(times[0]), 
+                            'duration': self._time_to_sec(times[1]) - self._time_to_sec(times[0])
                         })
         except: pass
         return segments
@@ -83,15 +77,12 @@ class PitchAnimationGenerator:
     def _parse_pitch_srt(self):
         pitch_info = []
         try:
-            if not os.path.exists(self.pitch_srt_path): return []
             with open(self.pitch_srt_path, 'r', encoding='utf-8') as f:
                 content = f.read().strip().split('\n\n')
                 for block in content:
                     lines = [l.strip() for l in block.split('\n') if l.strip()]
                     if len(lines) >= 3:
-                        t_range = lines[1]
-                        if ' --> ' not in t_range: continue
-                        times = t_range.split(' --> ')
+                        times = lines[1].split(' --> ')
                         orig = re.search(r'原唱: ([A-G]#?\d+)', lines[2])
                         cov = re.search(r'翻唱: ([A-G]#?\d+)', lines[2])
                         if orig and cov:
@@ -108,9 +99,8 @@ class PitchAnimationGenerator:
         for seg in self.summary_segments:
             seg['matched_pitches'] = [
                 p for p in self.all_pitch_info 
-                if 'end_sec' in p and \
-                p['start_sec'] >= seg['start_sec'] - 0.2 and \
-                p['end_sec'] <= seg['end_sec'] + 0.5
+                if 'end_sec' in p and p['start_sec'] >= seg['start_sec'] - 0.2 
+                and p['end_sec'] <= seg['start_sec'] + seg['duration'] + 0.5
             ]
         return self.summary_segments
 
@@ -118,20 +108,14 @@ class PitchAnimationGenerator:
         audio_map = {}
         if not os.path.exists(self.cover_audio_dir): return audio_map
         for f in os.listdir(self.cover_audio_dir):
-            match = re.search(r'句段(\d+)', f)
-            if match:
-                seq = int(match.group(1))
-                audio_map[seq] = os.path.join(self.cover_audio_dir, f)
+            m = re.search(r'句段(\d+)', f)
+            if m: audio_map[int(m.group(1))] = os.path.join(self.cover_audio_dir, f)
         return audio_map
 
     def generate_single_animation(self, segment):
         seq_num = segment['seq_num']
         duration = segment['duration']
-        
-        # --- 核心修改：不足1秒直接跳过 ---
-        if duration < 1.0:
-            print(f"⚠️ 句段 {seq_num} 长度不足1秒 ({duration:.2f}s)，跳过动画生成。")
-            return None
+        if duration < 1.0: return None # 跳过不足1秒
 
         pitches = segment['matched_pitches']
         if not pitches: return None
@@ -139,33 +123,27 @@ class PitchAnimationGenerator:
         all_midis = [p['orig_midi'] for p in pitches] + [p['cov_midi'] for p in pitches]
         y_min, y_max = int(min(all_midis)) - 1, int(max(all_midis)) + 1
 
-        # 调整布局，为图表上方留出文字空间
-        fig, ax = plt.subplots(figsize=(10, 6), dpi=self.dpi)
-        plt.subplots_adjust(top=0.85) # 留出顶部 15% 的高度
+        # 设置画布，增加右侧宽度比例以容纳外部图例
+        fig, ax = plt.subplots(figsize=(12, 6), dpi=self.dpi)
         
         ax.set_xlim(0, duration)
         ax.set_ylim(y_min, y_max)
         
-        # 纵轴音名展示
         y_ticks = np.arange(y_min, y_max + 1)
         ax.set_yticks(y_ticks)
         ax.set_yticklabels([self.midi_to_name[int(y)] for y in y_ticks], fontproperties=self.zh_font)
         
-        ax.set_title(f"句段 {seq_num} 音高对比轨迹", fontproperties=self.zh_font, pad=20)
-        ax.set_xlabel("时间 (秒)", fontproperties=self.zh_font)
-        ax.set_ylabel("音高", fontproperties=self.zh_font)
-        ax.grid(True, linestyle='--', alpha=0.4)
-
-        # --- 核心修改：信息面板移至图表外部 (Fig Title Area) ---
-        time_info = (f"句段：{seq_num}  |  起始时间：{segment['time_range'].split(' --> ')[0]}  "
-                     f"|  结束时间：{segment['time_range'].split(' --> ')[1]}")
+        # --- 核心修改：两行标题展示 ---
+        start_time = segment['time_range'].split(' --> ')[0]
+        end_time = segment['time_range'].split(' --> ')[1]
+        full_title = f"句段 {seq_num} 音高对比轨迹\n起止时间: {start_time} --> {end_time}"
+        ax.set_title(full_title, fontproperties=self.zh_font, fontsize=12, pad=15)
         
-        # 使用 fig.text 在画布顶部中央居中显示信息
-        fig.text(0.5, 0.92, time_info, fontproperties=self.zh_font, 
-                 fontsize=11, ha='center', fontweight='bold',
-                 bbox=dict(boxstyle='round,pad=0.5', facecolor='#f0f0f0', alpha=0.8, edgecolor='gray'))
+        ax.set_xlabel("相对时间 (秒)", fontproperties=self.zh_font)
+        ax.set_ylabel("音高名", fontproperties=self.zh_font)
+        ax.grid(True, linestyle='--', alpha=0.4)
 
-        # 1. 绘制原唱参考线 (蓝)
+        # 1. 绘制原唱参考线
         orig_x, orig_y = [], []
         for p in pitches:
             ps, pe = p['start_sec'] - segment['start_sec'], p['end_sec'] - segment['start_sec']
@@ -173,10 +151,15 @@ class PitchAnimationGenerator:
             orig_y.extend([p['orig_midi'], p['orig_midi'], None])
         ax.plot(orig_x, orig_y, color='#4169E1', lw=5, alpha=0.25, label='原唱参考')
 
-        # 2. 初始化翻唱线 (绿/红)
+        # 2. 初始化翻唱线
         match_line, = ax.plot([], [], color='#32CD32', lw=3, label='翻唱准确')
         mismatch_line, = ax.plot([], [], color='#DC143C', lw=3, label='翻唱走音')
-        ax.legend(prop=self.zh_font, loc='upper right', framealpha=0.5)
+        
+        # --- 核心修改：图例移至右侧外部 ---
+        ax.legend(prop=self.zh_font, loc='upper left', bbox_to_anchor=(1.02, 1), borderaxespad=0.)
+
+        # 调整整体布局，为右侧图例和顶部双行标题留空
+        plt.tight_layout(rect=[0, 0, 0.85, 1]) 
 
         frames = int(duration * self.fps)
         t_axis = np.linspace(0, duration, frames)
@@ -185,8 +168,7 @@ class PitchAnimationGenerator:
             curr_t = t_axis[frame]
             mx, my, ex, ey = [], [], [], []
             for p in pitches:
-                ps = p['start_sec'] - segment['start_sec']
-                pe = p['end_sec'] - segment['start_sec']
+                ps, pe = p['start_sec'] - segment['start_sec'], p['end_sec'] - segment['start_sec']
                 if ps > curr_t: continue
                 display_pe = min(pe, curr_t)
                 is_match = abs(p['orig_midi'] - p['cov_midi']) < 0.5
@@ -214,8 +196,7 @@ class PitchAnimationGenerator:
         try:
             video = VideoFileClip(video_p)
             audio = AudioFileClip(audio_p).subclip(0, video.duration)
-            final = video.set_audio(audio)
-            final.write_videofile(
+            video.set_audio(audio).write_videofile(
                 out_p, codec="libx264", audio_codec="aac", 
                 logger=None, threads=1, preset="ultrafast",
                 ffmpeg_params=["-pix_fmt", "yuv420p"]