From b19bd9a81a4d67920cc67236edb9755e6cdaed25 Mon Sep 17 00:00:00 2001
From: Dave Mateer <davemateer@gmail.com>
Date: Mon, 25 Apr 2022 16:14:59 +0100
Subject: [PATCH] auto

---
 .vscode/launch.json        |  9 +++++++++
 archivers/base_archiver.py |  5 -----
 auto_archive.py            | 20 ++++++++++++++------
 3 files changed, 23 insertions(+), 11 deletions(-)

diff --git a/.vscode/launch.json b/.vscode/launch.json
index fe359b1..a1bd561 100644
--- a/.vscode/launch.json
+++ b/.vscode/launch.json
@@ -19,6 +19,15 @@
 			"justMyCode": true,
 			"args": ["--sheet","Test Hashing CIR"]
 		},
+		{
+			"name": "Archive - Dearbhla",
+			"type": "python",
+			"request": "launch",
+			"program": "auto_archive.py",
+			"console": "integratedTerminal",
+			"justMyCode": true,
+			"args": ["--sheet","Archive - Dearbhla"]
+		},
 		{
 			"name": "Kayleigh - test - DM",
 			"type": "python",
diff --git a/archivers/base_archiver.py b/archivers/base_archiver.py
index af1fca6..d30f832 100644
--- a/archivers/base_archiver.py
+++ b/archivers/base_archiver.py
@@ -127,12 +127,10 @@ class Archiver(ABC):
         return hash.hexdigest()
 
     def get_screenshot(self, url):
-        logger.debug(f'In get_screenshot for {url}')
         key = self.get_key(urlparse(url).path.replace(
             "/", "_") + datetime.datetime.utcnow().isoformat().replace(" ", "_") + ".png")
         filename = 'tmp/' + key
 
-
         # DM - Accept cookies popup dismiss for ytdlp video
         if 'facebook.com' in url:
             try:
@@ -141,11 +139,8 @@ class Archiver(ABC):
             except:
                 logger.error('Failed on fb accept cookies')
         
-        logger.debug(f'get_screenshot: Requesting url')
-
         self.driver.get(url)
 
-        logger.debug(f'get_screenshot: Back from request')
         time.sleep(6)
 
         self.driver.save_screenshot(filename)
diff --git a/auto_archive.py b/auto_archive.py
index 87a964a..31fedf3 100644
--- a/auto_archive.py
+++ b/auto_archive.py
@@ -15,8 +15,10 @@ from utils import GWorksheet, mkdir_if_not_exists
 
 import sys
 
-load_dotenv()
+logger.add("trace.log", level="TRACE")
+logger.add("warnings.log", level="WARNING")
 
+load_dotenv()
 
 def update_sheet(gw, row, result: archivers.ArchiveResult):
     cell_updates = []
@@ -67,6 +69,9 @@ def process_sheet(sheet, header=1, columns=GWorksheet.COLUMN_NAMES):
     gc = gspread.service_account(filename='service_account.json')
     sh = gc.open(sheet)
 
+    # DM test raise error for decorator to catch
+    # raise ValueError('A very specific bad thing happened.')
+
     s3_config = S3Config(
         bucket=os.getenv('DO_BUCKET'),
         region=os.getenv('DO_SPACES_REGION'),
@@ -88,7 +93,9 @@ def process_sheet(sheet, header=1, columns=GWorksheet.COLUMN_NAMES):
 
     # loop through worksheets to check
     for ii, wks in enumerate(sh.worksheets()):
-        logger.info(f'Opening worksheet {ii}: "{wks.title}" header={header}')
+        # logger.info(f'Opening worksheet {ii}: "{wks.title}" header={header}')
+        # DM take " out of log message and clarify ii
+        logger.info(f'Opening worksheet ii={ii}: {wks.title} header={header}')
         gw = GWorksheet(wks, header_row=header, columns=columns)
 
         if not gw.col_exists('url'):
@@ -134,6 +141,7 @@ def process_sheet(sheet, header=1, columns=GWorksheet.COLUMN_NAMES):
                         result = archiver.download(url, check_if_exists=True)
                     except Exception as e:
                         result = False
+                        # DM loguru writes traceback to files so this traceback may be superfluous
                         logger.error(f'Got unexpected error in row {row} with archiver {archiver} for url {url}: {e}\n{traceback.format_exc()}')
 
                     if result:
@@ -166,10 +174,11 @@ def process_sheet(sheet, header=1, columns=GWorksheet.COLUMN_NAMES):
         logger.success(f'Finshed worksheet {wks.title}')
     driver.quit()
 
-
+@logger.catch
 def main():
-
-    print(sys.argv[1:])
+    # DM don't want to use print anymore
+    # print(sys.argv[1:])
+    logger.info(f'Passed args:{sys.argv}')
 
     parser = argparse.ArgumentParser(
         description='Automatically archive social media videos from a Google Sheets document')
@@ -192,4 +201,3 @@ def main():
 
 if __name__ == '__main__':
     main()
-    logger.success("finished")