353 lines
12 KiB
Python
353 lines
12 KiB
Python
import requests
|
|
import json
|
|
from datetime import datetime, timedelta
|
|
from openpyxl import Workbook
|
|
from openpyxl.styles import Font, PatternFill, Alignment, Border, Side, numbers
|
|
from openpyxl.utils import get_column_letter
|
|
|
|
def get_time_range(days_ago=1):
|
|
end_time = datetime.now()
|
|
start_time = end_time - timedelta(days=days_ago)
|
|
return {
|
|
'start_time': start_time.strftime('%Y-%m-%d 00:00'),
|
|
'end_time': end_time.strftime('%Y-%m-%d 23:59')
|
|
}
|
|
|
|
cookies = {
|
|
'ut_user_id': 'null',
|
|
'ut_global_id': '%22ca426419-74d7-4dc8-bc5c-c866b096b297%22',
|
|
'_gcl_au': '1.1.505382516.1778740165',
|
|
'_ga': 'GA1.1.968517445.1778741596',
|
|
'sensorsdata2015jssdkcross': '%7B%22distinct_id%22%3A%2219e2542d1f1efd-0b7accb742cea4-4c657b58-2073600-19e2542d1f21d61%22%2C%22%24device_id%22%3A%2219e2542d1f1efd-0b7accb742cea4-4c657b58-2073600-19e2542d1f21d61%22%2C%22props%22%3A%7B%7D%7D',
|
|
'Qs_lvt_102458': '1778741596%2C1779092047',
|
|
'Hm_lvt_85cdbdd6ba7f014cd503e9f1cd5e5ba0': '1778741596,1779092047',
|
|
'Qs_pv_102458': '4477521906714103000%2C2213764348932670000%2C1655296003018746400%2C2955048170989231600%2C930806901093578800',
|
|
'_ga_WPQK651LHJ': 'GS2.1.s1779095976$o3$g0$t1779095976$j60$l0$h0',
|
|
}
|
|
|
|
headers = {
|
|
'Accept': '*/*',
|
|
'Accept-Language': 'zh-CN,zh;q=0.9,en;q=0.8,en-GB;q=0.7,en-US;q=0.6',
|
|
'Connection': 'keep-alive',
|
|
'Referer': 'https://agent.udesk.cn/app/a46947ea-c1bf-4884-b6ba-9d7be32e18c4/logs',
|
|
'Sec-Fetch-Dest': 'empty',
|
|
'Sec-Fetch-Mode': 'cors',
|
|
'Sec-Fetch-Site': 'same-origin',
|
|
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/148.0.0.0 Safari/537.36 Edg/148.0.0.0',
|
|
'authorization': 'Bearer eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJ1c2VyX2lkIjoiNWQzNjMzZjMtNTIzZS00NTRjLTgyYzgtMTczMjc2YjVjMjg0IiwiZXhwIjoxNzc5MjYwNjU4LCJpc3MiOiJTRUxGX0hPU1RFRCIsInN1YiI6IkNvbnNvbGUgQVBJIFBhc3Nwb3J0In0.RseHavVdtiKHMKB2Mm8WVm4KwhiUZKmG93TiMIKuwtQ',
|
|
'content-type': 'application/json',
|
|
'sec-ch-ua': '"Chromium";v="148", "Microsoft Edge";v="148", "Not/A)Brand";v="99"',
|
|
'sec-ch-ua-mobile': '?0',
|
|
'sec-ch-ua-platform': '"Windows"',
|
|
}
|
|
|
|
app_id = 'a46947ea-c1bf-4884-b6ba-9d7be32e18c4'
|
|
|
|
def fetch_nodes(wf_run_id):
|
|
if not wf_run_id:
|
|
return []
|
|
try:
|
|
response = requests.get(
|
|
f'https://agent.udesk.cn/api/backend/new/apps/{app_id}/wfRuns/{wf_run_id}/nodeExecutions',
|
|
cookies=cookies,
|
|
headers=headers,
|
|
)
|
|
data = response.json()
|
|
return data.get('data', {}).get('list', [])
|
|
except Exception as e:
|
|
print(f'获取节点执行记录失败 wf_run_id={wf_run_id}: {e}')
|
|
return []
|
|
|
|
def safe_json_dumps(obj):
|
|
if obj is None:
|
|
return ''
|
|
try:
|
|
return json.dumps(obj, ensure_ascii=False)
|
|
except Exception:
|
|
return str(obj)
|
|
|
|
def safe_str(val):
|
|
if val is None:
|
|
return ''
|
|
return str(val)
|
|
|
|
def ts_to_datetime(ts):
|
|
try:
|
|
return datetime.fromtimestamp(int(ts)).strftime('%Y-%m-%d %H:%M:%S')
|
|
except Exception:
|
|
return ''
|
|
|
|
KEY_NODE_TITLES = ['大模型', '单次反思', '大模型二次生成']
|
|
|
|
time_range = get_time_range(days_ago=1)
|
|
|
|
params = {
|
|
'page_number': '1',
|
|
'page_size': '10',
|
|
'start_time': time_range['start_time'],
|
|
'end_time': time_range['end_time'],
|
|
'order_by': '-created_at',
|
|
'status': 'all',
|
|
}
|
|
|
|
response = requests.get(
|
|
f'https://agent.udesk.cn/api/backend/apps/{app_id}/log/chat/pages',
|
|
params=params,
|
|
cookies=cookies,
|
|
headers=headers,
|
|
)
|
|
data = response.json()
|
|
total = data.get('data', {}).get('total', 0)
|
|
total_page = max(total // 10 + (1 if total % 10 else 0), 1)
|
|
print(f'共 {total} 条对话记录，{total_page} 页，开始拉取...')
|
|
|
|
chart_list = []
|
|
for page in range(1, total_page + 1):
|
|
params['page_number'] = str(page)
|
|
response = requests.get(
|
|
f'https://agent.udesk.cn/api/backend/apps/{app_id}/log/chat/pages',
|
|
params=params,
|
|
cookies=cookies,
|
|
headers=headers,
|
|
)
|
|
page_data = response.json().get('data', {}).get('list', [])
|
|
chart_list.extend(page_data)
|
|
print(f' 第{page}/{total_page}页，获取 {len(page_data)} 条')
|
|
|
|
print(f'\n对话列表获取完成，共 {len(chart_list)} 条')
|
|
|
|
records = []
|
|
|
|
def fetch_all_messages(chat_log_id):
|
|
all_messages = []
|
|
fid = None
|
|
page_size = 10
|
|
|
|
while True:
|
|
try:
|
|
c_params = {'id': chat_log_id, 'page_size': str(page_size)}
|
|
if fid:
|
|
c_params['fid'] = fid
|
|
|
|
c_resp = requests.get(
|
|
f'https://agent.udesk.cn/api/backend/apps/{app_id}/chatMessage',
|
|
params=c_params,
|
|
cookies=cookies,
|
|
headers=headers,
|
|
timeout=30
|
|
)
|
|
|
|
data = c_resp.json()
|
|
messages = data.get('data', {}).get('list', [])
|
|
|
|
if not messages:
|
|
break
|
|
|
|
all_messages.extend(messages)
|
|
|
|
if len(messages) < page_size:
|
|
break
|
|
|
|
fid = messages[0].get('id')
|
|
if not fid:
|
|
break
|
|
|
|
except Exception as e:
|
|
break
|
|
|
|
return all_messages
|
|
|
|
for idx, chart in enumerate(chart_list):
|
|
chat_log_id = chart.get('chat_log_id', '')
|
|
chat_title = chart.get('title') or chart.get('chat_log_name') or ''
|
|
created_time = str(chart.get('created_time', ''))
|
|
|
|
try:
|
|
messages = fetch_all_messages(chat_log_id)
|
|
print(f' [{idx+1}] 获取对话消息 {len(messages)} 条')
|
|
except Exception as e:
|
|
print(f' [{idx+1}] 获取对话消息失败 chat_log_id={chat_log_id}: {e}')
|
|
messages = []
|
|
|
|
for msg in messages:
|
|
wf_run_id = msg.get('workflow_run_id', '')
|
|
if not wf_run_id:
|
|
continue
|
|
nodes = fetch_nodes(wf_run_id)
|
|
|
|
user_query = ''
|
|
for n in nodes:
|
|
if n.get('node_type') == 'start' and n.get('outputs', {}).get('sys.query'):
|
|
user_query = n['outputs']['sys.query']
|
|
break
|
|
|
|
key_outputs = {}
|
|
final_answer = ''
|
|
for n in nodes:
|
|
title = n.get('title', '')
|
|
if title in KEY_NODE_TITLES:
|
|
key_outputs[title] = (n.get('outputs') or {}).get('text', '')
|
|
if n.get('node_type') == 'answer':
|
|
final_answer = (n.get('outputs') or {}).get('answer', '')
|
|
|
|
records.append({
|
|
'chat_log_id': chat_log_id,
|
|
'chat_title': chat_title,
|
|
'created_time': created_time,
|
|
'user_query': user_query,
|
|
'wf_run_id': wf_run_id,
|
|
'nodes': nodes,
|
|
'key_outputs': key_outputs,
|
|
'final_answer': final_answer,
|
|
})
|
|
|
|
print(f' [{idx+1}/{len(chart_list)}] {chat_log_id[:8]}... 节点数:{len(nodes)}')
|
|
|
|
print(f'\n数据拉取完成，共 {len(records)} 条有效记录，开始生成Excel...')
|
|
|
|
wb = Workbook()
|
|
|
|
ws1 = wb.active
|
|
ws1.title = '关键节点审核'
|
|
|
|
header_font = Font(name='微软雅黑', bold=True, size=11, color='FFFFFF')
|
|
header_fill = PatternFill(start_color='4472C4', end_color='4472C4', fill_type='solid')
|
|
key_fill = PatternFill(start_color='FFF2CC', end_color='FFF2CC', fill_type='solid')
|
|
data_font = Font(name='微软雅黑', size=10)
|
|
thin_border = Border(
|
|
left=Side(style='thin'),
|
|
right=Side(style='thin'),
|
|
top=Side(style='thin'),
|
|
bottom=Side(style='thin'),
|
|
)
|
|
wrap_alignment = Alignment(horizontal='left', vertical='top', wrap_text=True)
|
|
center_alignment = Alignment(horizontal='center', vertical='center', wrap_text=True)
|
|
|
|
sheet1_headers = [
|
|
'序号', '对话ID', '会话标题', '对话时间',
|
|
'用户问题',
|
|
'大模型_输出', '单次反思_输出', '大模型二次生成_输出',
|
|
'最终回答', '人工审核结果', '备注'
|
|
]
|
|
|
|
for col_idx, header in enumerate(sheet1_headers, 1):
|
|
cell = ws1.cell(row=1, column=col_idx, value=header)
|
|
cell.font = header_font
|
|
cell.fill = header_fill
|
|
cell.alignment = center_alignment
|
|
cell.border = thin_border
|
|
|
|
key_col_indices = [col_idx for col_idx, h in enumerate(sheet1_headers, 1) if h in [
|
|
'用户问题', '大模型_输出', '单次反思_输出', '大模型二次生成_输出', '最终回答'
|
|
]]
|
|
audit_col = sheet1_headers.index('人工审核结果') + 1
|
|
|
|
for row_idx, rec in enumerate(records):
|
|
r = row_idx + 2
|
|
values = [
|
|
row_idx + 1,
|
|
rec['chat_log_id'],
|
|
rec['chat_title'],
|
|
rec['created_time'],
|
|
rec['user_query'],
|
|
rec['key_outputs'].get('大模型', ''),
|
|
rec['key_outputs'].get('单次反思', ''),
|
|
rec['key_outputs'].get('大模型二次生成', ''),
|
|
rec['final_answer'],
|
|
'',
|
|
'',
|
|
]
|
|
for col_idx, val in enumerate(values, 1):
|
|
cell = ws1.cell(row=r, column=col_idx, value=val)
|
|
cell.font = data_font
|
|
cell.alignment = wrap_alignment
|
|
cell.border = thin_border
|
|
if col_idx in key_col_indices:
|
|
cell.fill = key_fill
|
|
|
|
ws1.column_dimensions['A'].width = 6
|
|
ws1.column_dimensions['B'].width = 38
|
|
ws1.column_dimensions['C'].width = 20
|
|
ws1.column_dimensions['D'].width = 20
|
|
ws1.column_dimensions['E'].width = 40
|
|
ws1.column_dimensions['F'].width = 50
|
|
ws1.column_dimensions['G'].width = 50
|
|
ws1.column_dimensions['H'].width = 50
|
|
ws1.column_dimensions['I'].width = 50
|
|
ws1.column_dimensions['J'].width = 15
|
|
ws1.column_dimensions['K'].width = 20
|
|
|
|
ws1.freeze_panes = 'E2'
|
|
ws1.auto_filter.ref = f'A1:K{len(records) + 1}'
|
|
|
|
ws2 = wb.create_sheet('全流程日志明细')
|
|
|
|
sheet2_headers = [
|
|
'序号', '对话ID', '用户问题', '工作流运行ID',
|
|
'节点序号', '节点类型', '节点标题', '节点状态',
|
|
'执行开始时间', '执行结束时间',
|
|
'节点输入(JSON)', '节点输出(JSON)', '错误信息'
|
|
]
|
|
|
|
for col_idx, header in enumerate(sheet2_headers, 1):
|
|
cell = ws2.cell(row=1, column=col_idx, value=header)
|
|
cell.font = header_font
|
|
cell.fill = PatternFill(start_color='548235', end_color='548235', fill_type='solid')
|
|
cell.alignment = center_alignment
|
|
cell.border = thin_border
|
|
|
|
log_row = 2
|
|
log_seq = 0
|
|
for rec in records:
|
|
for node in rec['nodes']:
|
|
log_seq += 1
|
|
created_at = node.get('created_at') or node.get('started_at', 0)
|
|
finished_at = node.get('finished_at', 0)
|
|
values = [
|
|
log_seq,
|
|
rec['chat_log_id'],
|
|
rec['user_query'],
|
|
rec['wf_run_id'],
|
|
node.get('index', ''),
|
|
node.get('node_type', ''),
|
|
node.get('title', ''),
|
|
node.get('status', ''),
|
|
ts_to_datetime(created_at),
|
|
ts_to_datetime(finished_at),
|
|
safe_json_dumps(node.get('inputs')),
|
|
safe_json_dumps(node.get('outputs')),
|
|
safe_str(node.get('error', '')),
|
|
]
|
|
for col_idx, val in enumerate(values, 1):
|
|
cell = ws2.cell(row=log_row, column=col_idx, value=val)
|
|
cell.font = data_font
|
|
cell.alignment = wrap_alignment
|
|
cell.border = thin_border
|
|
if node.get('status') == 'failed':
|
|
cell.fill = PatternFill(start_color='FFC7CE', end_color='FFC7CE', fill_type='solid')
|
|
log_row += 1
|
|
|
|
ws2.column_dimensions['A'].width = 6
|
|
ws2.column_dimensions['B'].width = 38
|
|
ws2.column_dimensions['C'].width = 40
|
|
ws2.column_dimensions['D'].width = 38
|
|
ws2.column_dimensions['E'].width = 8
|
|
ws2.column_dimensions['F'].width = 22
|
|
ws2.column_dimensions['G'].width = 20
|
|
ws2.column_dimensions['H'].width = 12
|
|
ws2.column_dimensions['I'].width = 20
|
|
ws2.column_dimensions['J'].width = 20
|
|
ws2.column_dimensions['K'].width = 60
|
|
ws2.column_dimensions['L'].width = 60
|
|
ws2.column_dimensions['M'].width = 30
|
|
|
|
ws2.freeze_panes = 'F2'
|
|
ws2.auto_filter.ref = f'A1:M{log_row - 1}'
|
|
|
|
output_file = f'智能客服节点审核_{datetime.now().strftime("%Y%m%d_%H%M%S")}.xlsx'
|
|
output_path = f'd:/Idea Project/F6+宜搭+其它(1)/张阳脚本/udesk/{output_file}'
|
|
wb.save(output_path)
|
|
print(f'\nExcel已生成: {output_path}')
|
|
print(f' Sheet1「关键节点审核」: {len(records)} 条对话')
|
|
print(f' Sheet2「全流程日志明细」: {log_seq} 条节点执行记录')
|