ArchiveTeam/CommonCrawl_URL_Processor
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Update warc_wat_url_processor.py

Browse Source
This commit is contained in:
datechnoman 2024-01-28 11:23:27 +00:00
parent b6a9c68140
commit 54747b64f6
1 changed files with 26 additions and 19 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

19
warc_wat_url_processor.py
View File

@ -42,7 +42,7 @@ def extract_urls_from_file(file_path):
return urls
def process_file(file_path):
def process_file(file_path, error_log):
try:
print(f"Processing file: {file_path}")
@ -64,13 +64,15 @@ def process_file(file_path):
try:
# Run the compression command synchronously and wait for it to complete
result = subprocess.run(command_compress, shell=True, check=True)
result = subprocess.run(command_compress, shell=True, check=True, stderr=subprocess.PIPE, text=True)
if result.returncode == 0:
print(f"Compressed file saved as '{compressed_file_path}'")
else:
print(f"Compression failed for '{output_file_path}'")
error_log.write(f"Compression failed for '{output_file_path}': {result.stderr}\n")
except subprocess.CalledProcessError as compression_error:
print(f"Compression failed for '{output_file_path}': {compression_error}")
error_log.write(f"Compression failed for '{output_file_path}': {compression_error}\n")
# Remove the original gzipped file
os.remove(file_path)
@ -95,19 +97,22 @@ def process_file(file_path):
except Exception as e:
print(f"Error during processing {file_path}: {e}")
traceback.print_exc()
error_log.write(f"Error during processing {file_path}: {e}\n")
def download_and_process_file(url):
def download_and_process_file(url, error_log):
try:
command = f'axel -n 3 {url}'
result = subprocess.run(command, shell=True, check=True)
result = subprocess.run(command, shell=True, check=True, stderr=subprocess.PIPE, text=True)
if result.returncode == 0:
file_path = os.path.join(os.getcwd(), os.path.basename(url))
process_file(file_path)
process_file(file_path, error_log)
else:
print(f"Download failed for {url}")
error_log.write(f"Download failed for {url}\n")
except Exception as e:
print(f"Error during download and processing {url}: {e}")
error_log.write(f"Error during download and processing {url}: {e}\n")
def main():
check_disk_space('/dev/sda1')
@ -119,9 +124,10 @@ def main():
download_concurrency_level = 40
with open('error.log', 'a') as error_log:
with ProcessPoolExecutor(max_workers=download_concurrency_level) as executor:
print("Submitting tasks to the ProcessPoolExecutor...")
futures = [executor.submit(download_and_process_file, url) for url in urls]
futures = [executor.submit(download_and_process_file, url, error_log) for url in urls]
print(f"Submitted {len(futures)} tasks.")
print("Waiting for tasks to complete...")
@ -135,6 +141,7 @@ def main():
# Process the result if needed
except Exception as e:
print(f"Error in processing future: {e}")
error_log.write(f"Error in processing future: {e}\n")
if __name__ == "__main__":
main()