Crate dyer[−][src]
dyer is designed for reliable, flexible and fast web-crawling, providing some high-level, comprehensive features without compromising speed.
By means of event-driven, non-blocking I/O tokio and powerful, reliable Rust programming
language, inspired by scrapy, dyer
provides some high-level features:
- asynchronous, concurrent streamimg and I/O, make the best of thread pool, network, and system resource.
- event-driven, once you set the initials and recursive generator of Task,
dyer
will handle the rest of it. - user-friendly, considering the philosophy of rust programming language, more source code,
proper architecture may set up yourself in a dilemma when efficiency and learning cost are taken
into consideration.
dyer
offers high-level,flexible wrappers and APIs what does a lot for you.
Get Started by installing dyer-cli and looking over the examples.
Main Functionalities
Dyer
is newly developed rust library, and has achieved some basic functionalities for
building a crawer, web service and data processing. Nevertheless, it can tackle most common problems you meet.
Real-browser Customization
For each feeded Task, dyer
will fake a Profile and combines them into a Request to
satisfy the requirement of the target site. By means of powerful ffi
interface of and web
assemble of rust, intergration is not hard.
Signal Handling
Think about a scenario that errors, bugs and unexpected accidents are found when your app is running, what would you
do? Stop the app, the entire program and the data are corupted. Nope, the result is not
reliable. dyer
will backup your history between certain gap, resumption is at your choice.
Run-time Control
Each Task and each Profile is scheduled with some gap, has a time stamp for validation,
only the expired can be feeded to engine of dyer
. Nevertheless dyer
will limit the
Request sent to poll, the Profile to make, Task to load or store and so on see AppArg
.
Quick Start
dyer-cli is a handy tool for your easy and fast use of dyer, and recommanded to intergrate
with other dependencies. with rustup
and cargo
installed, the following code helps you get
the tool:
cargo install dyer-cli
Once installed, run dyer-cli
in your terminal or cmd prompt, it prints some info like
Handy tool for dyer
then dyer-cli is successfully installed.
Create A Project
Dyer-cli generates a template that contains many useful instances and instructions when using dyer with following code:
dyer-cli new myproject
It will create a project called myproject
and the files layout displays:
|___Cargo.toml
|___Readme.md
|___data/
|___data/tasks/
|___src/
|___src/entity.rs
|___src/parser.rs
|___src/spider.rs
|___src/middleware.rs
|___src/main.rs
|___src/pipeline.rs
Main functionality of each file:
- the
entity.rs
contains entities/data structure to be used/collected - the
parser.rs
contains functions that extract entities from response - the
spider.rs
contains initial when opening and final things to do when closing - the
middleware.rs
contains Some middlewares that process data at runtime - the
pipeline.rs
contains entities manipulation including data-storage, displsying and so on - the
main.rs
combines all modules then build them up into a programm Cargo.toml
is the basic configuration of the projectREADME.md
contains some instructions of the projectdata
folder balance the app load when data in app exceeds, and backup app data at certain gap
It is believed that learning by example is the best, and some examples are provided to illustrate how to use.
Features to be added
- proxy support
- debugging support(not bad though for now)
- more signal support(Ctrl+c for now)
- autothrottling and more customized plugins support
- more to go
Problem And Feedback
It is, of course, probable that bugs and errors lie in somewhere, and defects may appear in an unexpected way, if you got any one, comments and suggestions are welcome, please new a issue in my github.
Modules
component | |
engine | |
plugin |
Macros
plug | Macro that make a plugin at ease including |