Merge pull request #63 from ClericPy/dev

1.7.2

requirements.txt

@@ -2,7 +2,7 @@ uniparser>=1.4.5
 fastapi
 uvicorn
 databases
-torequests>=5.0.3
+torequests>=5.0.4
 fire
 jinja2
 aiofiles

watchdogs/__init__.py

@@ -3,6 +3,6 @@
 from .config import Config
 from .main import init_app
 
-__version__ = '1.7.1'
+__version__ = '1.7.2'
 __all__ = ['Config', 'init_app']
 logging.getLogger('watchdogs').addHandler(logging.NullHandler())

watchdogs/app.py

@@ -19,7 +19,7 @@
 
 from . import __version__
 from .config import md5_checker
-from .crawler import crawl_once
+from .crawler import crawl_once, find_next_check_time
 from .models import Task, query_tasks, tasks
 from .settings import (Config, get_host_freq_list, refresh_token, release_app,
                        set_host_freq, setup_app)
@@ -119,6 +119,7 @@ async def index(request: Request, tag: str = ''):
         'custom_links': Config.custom_links,
         'callback_workers': Config.callback_handler.workers,
         'custom_tabs': Config.custom_tabs,
+        'work_hours_doc': find_next_check_time.__doc__,
     })
     init_vars_b64 = b64encode(init_vars_json.encode('u8')).decode('u8')
     kwargs['init_vars'] = init_vars_b64
@@ -204,7 +205,7 @@ async def force_crawl(task_name: str):
 async def load_tasks(
     task_name: Optional[str] = None,
     page: int = 1,
-    page_size: int = 30,
+    page_size: int = Config.default_page_size,
     order_by: str = 'last_change_time',
     sort: str = 'desc',
     tag: str = '',
@@ -469,8 +470,12 @@ async def post_lite(request: Request,
 
 
 @app.get("/lite")
-async def lite(request: Request, tag: str = '', sign: str = '', page: int = 1):
-    tasks, has_more = await query_tasks(tag=tag, page=page)
+async def lite(request: Request,
+               tag: str = '',
+               sign: str = '',
+               page: int = 1,
+               page_size: int = Config.default_page_size):
+    tasks, has_more = await query_tasks(tag=tag, page=page, page_size=page_size)
     now = datetime.now()
     for task in tasks:
         result = loads(task['latest_result'] or '{}')

watchdogs/background.py

@@ -6,7 +6,6 @@
 async def crawl_chunks(crawl_once):
     loop_num = 0
     while not Config.is_shutdown:
-        loop_num += 1
         has_more = await crawl_once()
         if isinstance(has_more, Exception):
             Config.logger.error(f'crawl_once error, {has_more!r}')
@@ -15,6 +14,7 @@ async def crawl_chunks(crawl_once):
             f'crawl_once finished, has_more: {has_more}, loop: {loop_num}')
         if not has_more:
             break
+        loop_num += 1
 
 
 async def background_loop(coro_funcs: list = None):

watchdogs/config.py

@@ -181,6 +181,7 @@ class Config:
     custom_tabs: List[Dict] = []
     COLLATION: str = None
     cookie_max_age = 86400 * 7
+    default_page_size = 15
 
     @classmethod
     def add_custom_tabs(cls, label, url, name=None, desc=None):

watchdogs/crawler.py

@@ -1,9 +1,11 @@
+# -*- coding: utf-8 -*-
+
 from asyncio import ensure_future, wait
 from datetime import datetime, timedelta
 from json import JSONDecodeError, dumps, loads
 from typing import Optional, Tuple
 
-from torequests.utils import ttime
+from torequests.utils import timeago, ttime
 from uniparser import Crawler, RuleNotFoundError
 
 from .config import Config
@@ -38,8 +40,7 @@ def kwargs(self):
 
 
 def find_next_check_time(
-        work_hours: str,
-        interval: int,
+        task: Task,
         now: Optional[datetime] = None,
 ) -> Tuple[bool, datetime]:
     '''
@@ -63,30 +64,50 @@ def find_next_check_time(
             %w==5|20, 24        means every Friday or everyday 20:00 ~ 23:59
             %w==5|%w==2         means every Friday or Tuesday
             %w!=6&%w!=0         means everyday except Saturday & Sunday.
+        5. Set a ensure change interval
+            > If work_hours string endswith `#` and `x` seconds, will check the next_change_time first.
+            > In other words, I am very sure that the interval between two changes is more than `x` seconds
+            > So the crawler of this task will not run until the time is `last_change_time + change_interval`
+            %w==5#86400        means every Friday if it didn't change within 1 day
+            0, 24#3600         means each hour if it didn't change within this hour. The task will only be crawled once if it has changed.
     '''
     # find the latest hour fit work_hours, if not exist, return next day 00:00
     now = now or datetime.now()
+    work_hours = task.work_hours or '0, 24'
+    if '#' in work_hours:
+        # check if changed
+        last_change_time = task.last_change_time or datetime.fromtimestamp(0)
+        # split work_hours and change_interval
+        work_hours, change_interval_str = work_hours.split('#')
+        change_interval = int(change_interval_str)
+        # not fit change interval, will wait for left seconds.
+        next_change_time = last_change_time + timedelta(seconds=change_interval)
+        if now < next_change_time:
+            Config.logger.info(
+                f'Task [{task.name}] has changed in {timeago(change_interval, accuracy=1, format=1, short_name=1)} ago.'
+            )
+            return False, next_change_time
 
-    ok = check_work_time(work_hours, now)
-    if ok:
-        # current time is ok, next_check_time is now+interval
-        next_check_time = now + timedelta(seconds=interval)
-        return ok, next_check_time
+    need_crawl = check_work_time(work_hours, now)
+    if need_crawl:
+        # current time is need_crawl, next_check_time is now+interval
+        next_check_time = now + timedelta(seconds=task.interval)
+        return need_crawl, next_check_time
     else:
-        # current time is not ok
+        # current time is not need_crawl
         next_check_time = now
-        # time machine to check time fast
+        # time machine to update next_check_time fast
         for _ in range(60):
-            # check next interval
-            next_check_time = next_check_time + timedelta(seconds=interval)
-            _ok = check_work_time(work_hours, next_check_time)
-            if _ok:
-                # current is still False, but next_check_time is True
+            # next interval
+            next_check_time = next_check_time + timedelta(seconds=task.interval)
+            _need_crawl = check_work_time(work_hours, next_check_time)
+            if _need_crawl:
+                # current time is still False, but next_check_time is True
                 break
-        return ok, next_check_time
+        return need_crawl, next_check_time
 
 
-async def crawl(task):
+async def crawl(task: Task):
     crawler: Crawler = Config.crawler
     logger = Config.logger
     logger.info(f'Start crawling: {task.name}')
@@ -103,13 +124,17 @@ async def crawl(task):
     else:
         if len(crawl_result) == 1:
             # chain result for __request__ which fetch a new request
-            result_list = get_watchdog_result(item=crawl_result.popitem()[1])
-            if result_list == {'text': 'text not found'}:
+            formated_result = get_watchdog_result(
+                item=crawl_result.popitem()[1])
+            if formated_result == {'text': 'text not found'}:
                 error = f'{task.name} text not found, crawl result given: {crawl_result}'
                 logger.error(error)
+                result_list = None
             else:
-                if not isinstance(result_list, list):
-                    result_list = [result_list]
+                if isinstance(formated_result, list):
+                    result_list = formated_result
+                else:
+                    result_list = [formated_result]
                 # use force crawl one web UI for more log
                 logger.info(f'{task.name} Crawl success: {result_list}'[:150])
         else:
@@ -141,12 +166,11 @@ async def _crawl_once(task_name: Optional[str] = None, chunk_size: int = 20):
     for _task in fetched_tasks:
         task = Task(**dict(_task))
         # check work hours
-        ok, next_check_time = find_next_check_time(task.work_hours or '0, 24',
-                                                   task.interval, now)
+        need_crawl, next_check_time = find_next_check_time(task, now)
         if task_name:
             # always crawl for given task_name
-            ok = True
-        if ok:
+            need_crawl = True
+        if need_crawl:
             t = ensure_future(crawl(task))
             # add task_name for logger
             setattr(t, 'task_name', task.name)
@@ -160,7 +184,7 @@ async def _crawl_once(task_name: Optional[str] = None, chunk_size: int = 20):
         # update task variable for callback
         task.__dict__.update(values)
         update_values.append(values)
-        if not ok:
+        if not need_crawl:
             logger.info(
                 f'Task [{task.name}] is not on work time, next_check_time reset to {next_check_time}'
             )

watchdogs/models.py

@@ -274,7 +274,7 @@ async def query_tasks(
         task_name: Optional[str] = None,
         task_id: Optional[int] = None,
         page: int = 1,
-        page_size: int = 30,
+        page_size: int = Config.default_page_size,
         order_by: str = 'last_change_time',
         sort: str = 'desc',
         tag: str = '',

watchdogs/static/css/watchdogs.css

@@ -79,6 +79,9 @@ p.custom_links {
 .el-popover {
     max-width: 50%;
 }
+.el-message-box.work_hours_doc{
+    width: 40%;
+}
 pre {
     word-wrap: break-word;
     white-space: pre-wrap;