Aget-rs - Fast Asynchronous Downloader with Rust.
aget-rs
is a fast asynchronous downloader wrote with Rust.
It requests a resource with a number of concurrent asynchronous request in a single thread.
Especially, the concurrent amount can be any positive number as your wish.
aget-rs
supports to download a HTTP/S link and a M3U8 video link.
You can download the last release from https://github.com/PeterDing/aget-rs/releases
We conside that there is a file to download. This file has 10MB. The server which hosts the file has been set a limit rate to 100KB/s, but no connecting count limit.
It will be easy to calculate the total costing time when we use 1, 10, 100 connections to request the file.
In the benchmark test, we use nginx
to simulate the environment where a limit rate is 100KB/s for downloading.
Following is the results of using curl
and aget-rs
. (For more details, you can find at here)
-
One connection using
curl
time curl http://localhost:9010/abc % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 100 10.0M 100 10.0M 0 0 100k 0 0:01:42 0:01:42 --:--:-- 103k real 1m42.147s user 0m0.021s sys 0m0.035s
time cost: 102s
10 connections using aget-rs
time ag http://localhost:9010/abc -s 10 -k 1m
File: abc
Length: 10.0M (10485760)
10.0M/10.0M 100.00% NaNG/s eta: 0s [==================================>]
real 0m10.016s
user 0m0.040s
sys 0m0.020s
time cost: 10s, 10 times faster than curl
100 connections using aget-rs
time ag http://localhost:9010/abc -s 100 -k 103k
File: abc
Length: 10.0M (10485760)
10.0M/10.0M 100.00% NaNG/s eta: 0s [==================================>]
real 0m2.016s
user 0m0.087s
sys 0m0.029s
time cost: 2s, 50 times faster than curl
Request a resource with default configuration
The default concurrent amount is 10
and chunk length is 1m
.
ag http://cdimage.ubuntu.com/ubuntu/releases/18.10/release/ubuntu-18.10-server-amd64.iso
Set concurrent amount and chunk length
Use -s
or --concurrency
to set the number of concurrent request.
Use -k
or --chunk-size
to set the chunk length of each request.
--chunk-size
takes a literal size description, example 1k
for one Kilobyte,
2m
for two Megabyte, 1g
for Gigabyte.
ag "url of resource" -s 20 -k 1m
Set a path for output
Use -o
or --out
to set the path.
If the argument is not gave, we take the last part of the url' path as the path.
ag "url of resource" -o /path/to/file
Set request headers
Use -H
to set headers.
ag "url of resource" -H "Cookie: key=value" -H "Accept: */*"
Set request method and data
Use -X
or --method
to set method for http, example, GET
, POST
.
The default method is GET
.
With a data, using -d
or --data
, example, a=b
ag "url of resource" -d "a=b"
Usage: ag [OPTIONS] <URL>
Arguments:
<URL>
Options:
-m, --method <METHOD> Request method, e.g. GET, POST [default: GET]
-H, --header <HEADER> Request headers, e.g. -H "User-Agent: aget"
-d, --data <DATA> Request with POST method with the data, e.g. -d "a=b"
--insecure Skip to verify the server's TLS certificate
-s, --concurrency <CONCURRENCY> The number of concurrency request [default: 10]
-k, --chunk-size <CHUNK_SIZE> The number ofinterval length of each concurrent request [default: '50m']
-t, --timeout <TIMEOUT> Timeout(seconds) of request [default: 60]
--dns-timeout <DNS_TIMEOUT> DNS Timeout(seconds) of request [default: 10]
--retries <RETRIES> The maximum times of retring [default: 5]
--retry-wait <RETRY_WAIT> The seconds between retries [default: 0]
--proxy <PROXY> [protocol://]host[:port] Use this proxy
--type <TYPE> Task type, auto/http/m3u8 [default: auto]
--debug Debug output. Print all trackback for debugging
--quiet Quiet mode. Don't show progress bar and task information. But still show the error information
-o, --out <OUT> The path of output for the request e.g. -o "/path/to/file"
-h, --help Print help
-V, --version Print version
Aget can be configured by a configuration file. The file locates at ~/.config/aget/config
.
Following options can be set. Aget uses these options as the defaults for each command.
headers = [["key", "value"], ...]
concurrency = ...
chunk_size = "..."
timeout = ...
dns_timeout = ...
retries = ...
retry_wait = ...
If the file does not exist, aget will use the default configuration.
headers = [["user-agent", "aget/version"]]
concurrency = 10
chunk_size = "50m"
timeout = 60
dns_timeout = 10
retries = 5
retry_wait = 0
from https://github.com/PeterDing/aget-rs
No comments:
Post a Comment