JustAnotherArchivist
73f35f5591
Fix infinite loop when file ends with something that is not a WARC record
2 anni fa
JustAnotherArchivist
06d60a798c
Bump read size
2 anni fa
JustAnotherArchivist
3e0b70be6b
Handle processes with too many open connections
2 anni fa
JustAnotherArchivist
df7b25c2db
Error on unknown options
2 anni fa
JustAnotherArchivist
4bd4f5a30c
Fix 'Argument list too long' error when using --urls-from-stdin with many URLs
2 anni fa
JustAnotherArchivist
e20d35a553
Fix crash on 429
2 anni fa
JustAnotherArchivist
cef61434a0
Add --urls-from-stdin
2 anni fa
JustAnotherArchivist
b5cf04947b
Add Wasabi
2 anni fa
JustAnotherArchivist
d2afd1309d
Add s3-bucket-find-direct-url
2 anni fa
JustAnotherArchivist
95988466ec
Make S3 response pattern matching more flexible (so it also works on Scaleway)
2 anni fa
JustAnotherArchivist
a9a03d3a00
Add urlsort
2 anni fa
JustAnotherArchivist
9798cc1188
Typo
2 anni fa
JustAnotherArchivist
d193637e5e
Add kill-connections
2 anni fa
JustAnotherArchivist
6cfe8e51ba
Make job a global variable in --pyfilter expressions so it can be used in genexps
2 anni fa
JustAnotherArchivist
a4627fa1c6
Queue derives with `ia tasks` instead of this manual curl rubbish
2 anni fa
JustAnotherArchivist
c68b310afc
Always print the parts value if there is an upload ID
Previously, parts wouldn't be printed if it was an empty list. This made resuming uploads that crashed in the first part harder than necessary.
2 anni fa
JustAnotherArchivist
fdc3c3d69e
Support float values for --partsize with M or G suffix
2 anni fa
JustAnotherArchivist
002c1eb7ae
Wait until item exists
IA doesn't immediately create the item on CreateMultipartUpload, so if it didn't already exist, UploadPart would fail for a while and we'd waste bandwidth.
2 anni fa
JustAnotherArchivist
142a5a9c49
Get rid of asyncio
No point in using it when it only delegates to a ThreadPoolExecutor anyway.
2 anni fa
JustAnotherArchivist
b6663ae731
Add concurrency
2 anni fa
JustAnotherArchivist
22f2e68356
Add JSONL output option for S3 listing
2 anni fa
JustAnotherArchivist
bfebe9a2a5
Fix only sending partial file contents on retries
2 anni fa
JustAnotherArchivist
39b3b7793a
Add support for IA_CONFIG_FILE environment variable
2 anni fa
JustAnotherArchivist
7ed2906dd2
Add progress bar
2 anni fa
JustAnotherArchivist
58f0f0f8d0
Fix being unable to resume an upload that crashed in the first part
2 anni fa
JustAnotherArchivist
74485c399b
Require decompressed WARCs with warc-tiny
2 anni fa
JustAnotherArchivist
e24790132e
Add at-tracker-sample-user-item-size
2 anni fa
JustAnotherArchivist
a14939b069
Add base64url
2 anni fa
JustAnotherArchivist
5c2ce7ec10
Add cdx-chunk
2 anni fa
JustAnotherArchivist
fe0b020352
Add support for reading from stdin
2 anni fa
JustAnotherArchivist
1010769c3c
Handle connection errors
2 anni fa
JustAnotherArchivist
1acdc88c81
Add ia-upload-stream
2 anni fa
JustAnotherArchivist
360c4d9371
Add youtube-extract-rapid
2 anni fa
JustAnotherArchivist
d07b5a7d09
Remove debugging prints
2 anni fa
JustAnotherArchivist
bf5e065a0f
Add URL/percent decoding tool
urldecode.c is entirely written by OrIdow6 except for one bug fix (char → uint8_t in the mallocs) and whitespace changes. The test suite is by JAA.
Co-authored-by: OrIdow6 <68304414+OrIdow6@users.noreply.github.com>
2 anni fa
JustAnotherArchivist
11485d9404
Add infrastructure for simple C-based tools
2 anni fa
JustAnotherArchivist
c50a8fd796
Fix 'Dictionary mismatch' error when very small dicts are used because the temporary file isn't written to disk before zstdcat gets executed
2 anni fa
JustAnotherArchivist
5bc3d4b020
Fix crash on an empty response
This check was a leftover from the resumeKey pagination, where empty responses are supposed to be impossible. With the page pagination, they are possible.
2 anni fa
JustAnotherArchivist
7f25c092d1
Catch other connection errors
2 anni fa
JustAnotherArchivist
f8352809f3
Handle ConnectionResetError
2 anni fa
JustAnotherArchivist
0b34268210
Catch socket.timeout, which is a separate exception class from TimeoutError before Python 3.10
2 anni fa
JustAnotherArchivist
0f7a2b32a3
Log number of results on a page
2 anni fa
JustAnotherArchivist
628aeb052f
Handle rate limiting
2 anni fa
JustAnotherArchivist
d3ea3ce8a0
Switch from urllib to http.client to reuse connections
2 anni fa
JustAnotherArchivist
8f7619ff3a
Add retries
2 anni fa
JustAnotherArchivist
f98fdd5f01
Fix printing HTTP response line to stdout instead of stderr
2 anni fa
JustAnotherArchivist
c9400ac46f
Fix recognition of command without optional parts
2 anni fa
JustAnotherArchivist
5ca15a7c94
Add concurrency support
The proper way to do that (with asyncio) is of course aiohttp. A major drawback of the implemented approach is that running tasks can't be cancelled in case of an error. However, it works with just the standard library, and that advantage outweighs the awkward error handling for now.
2 anni fa
JustAnotherArchivist
191948cf9d
Print number of modified records on requeueing
2 anni fa
JustAnotherArchivist
5121524f83
Log retrieval of showNumPages
2 anni fa