feat: update Python processors and add utility scripts
- Update ASR, face, OCR, pose processors - Add release pre-flight check script - Add synonym generation, chunk processing scripts - Add face recognition, stamp search utilities
This commit is contained in:
Warren
246
scripts/face_statistics_report.py
Normal file
246
scripts/face_statistics_report.py
Normal file
@@ -0,0 +1,246 @@
|
||||
#!/usr/bin/env python3
|
||||
"""
|
||||
人臉統計報告生成
|
||||
"""
|
||||
|
||||
import psycopg2
|
||||
import json
|
||||
from datetime import datetime
|
||||
import sys
|
||||
|
||||
|
||||
def get_face_statistics():
|
||||
"""獲取人臉統計數據"""
|
||||
conn = psycopg2.connect(
|
||||
host="localhost",
|
||||
port=5432,
|
||||
database="momentry",
|
||||
user="accusys",
|
||||
password="accusys",
|
||||
)
|
||||
|
||||
cursor = conn.cursor()
|
||||
|
||||
# 1. 總體統計
|
||||
cursor.execute("""
|
||||
SELECT
|
||||
COUNT(*) as total_faces,
|
||||
SUM(CASE WHEN attributes->>'gender' = 'male' THEN 1 ELSE 0 END) as male_count,
|
||||
SUM(CASE WHEN attributes->>'gender' = 'female' THEN 1 ELSE 0 END) as female_count,
|
||||
ROUND(AVG(CASE WHEN attributes->>'age' ~ '^[0-9]+$' THEN (attributes->>'age')::numeric ELSE NULL END)::numeric, 1) as avg_age,
|
||||
MIN(CASE WHEN attributes->>'age' ~ '^[0-9]+$' THEN (attributes->>'age')::numeric ELSE NULL END) as min_age,
|
||||
MAX(CASE WHEN attributes->>'age' ~ '^[0-9]+$' THEN (attributes->>'age')::numeric ELSE NULL END) as max_age
|
||||
FROM face_detections
|
||||
""")
|
||||
|
||||
total_stats = cursor.fetchone()
|
||||
|
||||
# 2. 按視頻統計
|
||||
cursor.execute("""
|
||||
SELECT
|
||||
video_uuid,
|
||||
COUNT(*) as total_faces,
|
||||
SUM(CASE WHEN attributes->>'gender' = 'male' THEN 1 ELSE 0 END) as male_count,
|
||||
SUM(CASE WHEN attributes->>'gender' = 'female' THEN 1 ELSE 0 END) as female_count,
|
||||
ROUND(AVG(CASE WHEN attributes->>'age' ~ '^[0-9]+$' THEN (attributes->>'age')::numeric ELSE NULL END)::numeric, 1) as avg_age
|
||||
FROM face_detections
|
||||
GROUP BY video_uuid
|
||||
ORDER BY total_faces DESC
|
||||
""")
|
||||
|
||||
video_stats = cursor.fetchall()
|
||||
|
||||
# 3. 年齡性別分布
|
||||
cursor.execute("""
|
||||
WITH age_groups AS (
|
||||
SELECT
|
||||
CASE
|
||||
WHEN (attributes->>'age')::numeric < 20 THEN '10-19'
|
||||
WHEN (attributes->>'age')::numeric < 30 THEN '20-29'
|
||||
WHEN (attributes->>'age')::numeric < 40 THEN '30-39'
|
||||
WHEN (attributes->>'age')::numeric < 50 THEN '40-49'
|
||||
WHEN (attributes->>'age')::numeric < 60 THEN '50-59'
|
||||
ELSE '60+'
|
||||
END as age_group,
|
||||
attributes->>'gender' as gender
|
||||
FROM face_detections
|
||||
WHERE attributes->>'gender' IN ('male', 'female')
|
||||
AND attributes->>'age' ~ '^[0-9]+$'
|
||||
)
|
||||
SELECT
|
||||
age_group,
|
||||
gender,
|
||||
COUNT(*) as count
|
||||
FROM age_groups
|
||||
GROUP BY age_group, gender
|
||||
ORDER BY
|
||||
CASE age_group
|
||||
WHEN '10-19' THEN 1
|
||||
WHEN '20-29' THEN 2
|
||||
WHEN '30-39' THEN 3
|
||||
WHEN '40-49' THEN 4
|
||||
WHEN '50-59' THEN 5
|
||||
ELSE 6
|
||||
END,
|
||||
gender DESC
|
||||
""")
|
||||
|
||||
age_gender_dist = cursor.fetchall()
|
||||
|
||||
# 4. 置信度統計
|
||||
cursor.execute("""
|
||||
SELECT
|
||||
ROUND(AVG(confidence)::numeric, 3) as avg_confidence,
|
||||
MIN(confidence) as min_confidence,
|
||||
MAX(confidence) as max_confidence,
|
||||
COUNT(CASE WHEN confidence >= 0.8 THEN 1 END) as high_confidence,
|
||||
COUNT(CASE WHEN confidence >= 0.6 AND confidence < 0.8 THEN 1 END) as medium_confidence,
|
||||
COUNT(CASE WHEN confidence < 0.6 THEN 1 END) as low_confidence
|
||||
FROM face_detections
|
||||
""")
|
||||
|
||||
confidence_stats = cursor.fetchone()
|
||||
|
||||
# 5. 時間分布
|
||||
cursor.execute("""
|
||||
SELECT
|
||||
FLOOR(timestamp_secs / 60) * 60 as minute_mark,
|
||||
COUNT(*) as faces_in_minute,
|
||||
SUM(CASE WHEN attributes->>'gender' = 'male' THEN 1 ELSE 0 END) as males_in_minute,
|
||||
SUM(CASE WHEN attributes->>'gender' = 'female' THEN 1 ELSE 0 END) as females_in_minute
|
||||
FROM face_detections
|
||||
GROUP BY FLOOR(timestamp_secs / 60) * 60
|
||||
ORDER BY minute_mark
|
||||
""")
|
||||
|
||||
time_dist = cursor.fetchall()
|
||||
|
||||
cursor.close()
|
||||
conn.close()
|
||||
|
||||
return {
|
||||
"total_stats": total_stats,
|
||||
"video_stats": video_stats,
|
||||
"age_gender_dist": age_gender_dist,
|
||||
"confidence_stats": confidence_stats,
|
||||
"time_dist": time_dist,
|
||||
}
|
||||
|
||||
|
||||
def generate_report(stats):
|
||||
"""生成統計報告"""
|
||||
report = []
|
||||
|
||||
report.append("=" * 70)
|
||||
report.append("人臉識別統計報告")
|
||||
report.append("=" * 70)
|
||||
report.append(f"生成時間: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}")
|
||||
report.append("")
|
||||
|
||||
# 總體統計
|
||||
total_stats = stats["total_stats"]
|
||||
report.append("📊 總體統計")
|
||||
report.append("-" * 40)
|
||||
report.append(f"總人臉數: {total_stats[0]}")
|
||||
report.append(
|
||||
f"男性: {total_stats[1]} ({total_stats[1] / total_stats[0] * 100:.1f}%)"
|
||||
)
|
||||
report.append(
|
||||
f"女性: {total_stats[2]} ({total_stats[2] / total_stats[0] * 100:.1f}%)"
|
||||
)
|
||||
report.append(f"平均年齡: {total_stats[3]} 歲")
|
||||
report.append(f"年齡範圍: {total_stats[4]} - {total_stats[5]} 歲")
|
||||
report.append("")
|
||||
|
||||
# 視頻統計
|
||||
report.append("🎬 視頻統計")
|
||||
report.append("-" * 40)
|
||||
for video in stats["video_stats"]:
|
||||
video_uuid, total, male, female, avg_age = video
|
||||
video_name = (
|
||||
"Old_Time_Movie_Show_-_Charade_1963.HD.mov"
|
||||
if video_uuid == "384b0ff44aaaa1f1"
|
||||
else "ExaSAN PCIe series"
|
||||
)
|
||||
report.append(f"視頻: {video_name}")
|
||||
report.append(f" UUID: {video_uuid}")
|
||||
report.append(f" 總人臉: {total}")
|
||||
report.append(f" 男性: {male} ({male / total * 100:.1f}%)")
|
||||
report.append(f" 女性: {female} ({female / total * 100:.1f}%)")
|
||||
report.append(f" 平均年齡: {avg_age} 歲")
|
||||
report.append("")
|
||||
|
||||
# 年齡性別分布
|
||||
report.append("👥 年齡性別分布")
|
||||
report.append("-" * 40)
|
||||
|
||||
# 創建分布表
|
||||
age_groups = {}
|
||||
for age_group, gender, count in stats["age_gender_dist"]:
|
||||
if age_group not in age_groups:
|
||||
age_groups[age_group] = {"male": 0, "female": 0}
|
||||
age_groups[age_group][gender] = count
|
||||
|
||||
for age_group in sorted(age_groups.keys(), key=lambda x: int(x.split("-")[0])):
|
||||
male = age_groups[age_group]["male"]
|
||||
female = age_groups[age_group]["female"]
|
||||
total = male + female
|
||||
if total > 0:
|
||||
report.append(f"{age_group}歲: {total}人 (男{male}/女{female})")
|
||||
|
||||
report.append("")
|
||||
|
||||
# 置信度統計
|
||||
conf_stats = stats["confidence_stats"]
|
||||
report.append("🎯 檢測置信度")
|
||||
report.append("-" * 40)
|
||||
report.append(f"平均置信度: {conf_stats[0]:.3f}")
|
||||
report.append(f"範圍: {conf_stats[1]:.3f} - {conf_stats[2]:.3f}")
|
||||
report.append(
|
||||
f"高置信度(≥0.8): {conf_stats[3]} ({conf_stats[3] / total_stats[0] * 100:.1f}%)"
|
||||
)
|
||||
report.append(
|
||||
f"中置信度(0.6-0.8): {conf_stats[4]} ({conf_stats[4] / total_stats[0] * 100:.1f}%)"
|
||||
)
|
||||
report.append(
|
||||
f"低置信度(<0.6): {conf_stats[5]} ({conf_stats[5] / total_stats[0] * 100:.1f}%)"
|
||||
)
|
||||
report.append("")
|
||||
|
||||
# 時間分布
|
||||
report.append("⏰ 時間分布 (每分鐘)")
|
||||
report.append("-" * 40)
|
||||
for minute_mark, total, male, female in stats["time_dist"]:
|
||||
minutes = int(minute_mark // 60)
|
||||
seconds = int(minute_mark % 60)
|
||||
report.append(f"{minutes:02d}:{seconds:02d} - {total}人 (男{male}/女{female})")
|
||||
|
||||
report.append("")
|
||||
report.append("=" * 70)
|
||||
|
||||
return "\n".join(report)
|
||||
|
||||
|
||||
def main():
|
||||
print("正在生成人臉統計報告...")
|
||||
|
||||
try:
|
||||
stats = get_face_statistics()
|
||||
report = generate_report(stats)
|
||||
|
||||
# 輸出到控制台
|
||||
print(report)
|
||||
|
||||
# 保存到文件
|
||||
with open("/tmp/face_statistics_report.txt", "w") as f:
|
||||
f.write(report)
|
||||
|
||||
print(f"\n報告已保存到: /tmp/face_statistics_report.txt")
|
||||
|
||||
except Exception as e:
|
||||
print(f"❌ 生成報告時出錯: {e}")
|
||||
sys.exit(1)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
Reference in New Issue
Block a user