Mirrors/the-algorithm
2
0
Fork 0
mirror of https://github.com/twitter/the-algorithm.git synced 2024-12-22 02:01:51 +01:00
Code Issues Packages Projects Releases Wiki Activity

Update README.md

Browse Source
This commit is contained in:
shungo 2023-04-01 19:09:34 +09:00 committed by GitHub
parent ec83d01dca
commit 637d0b9826
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
product-mixer/README.md
View File

@ -8,7 +8,7 @@ iterate on, and own product surface areas. It consists of:
- **Core Libraries:** A set of libraries that enable you to build execution pipelines out of
reusable components. You define your logic in small, well-defined, reusable components and focus
on expressing the business logic you want to have. Then you can define easy to understand pipelines
on expressing the business logic you want to have. Then you can define easy-to-understand pipelines
that compose your components. Product Mixer handles the execution and monitoring of your pipelines
allowing you to focus on what really matters, your business logic.
@ -22,7 +22,7 @@ iterate on, and own product surface areas. It consists of:
The bulk of a Product Mixer can be broken down into Pipelines and Components. Components allow you
to break business logic into separate, standardized, reusable, testable, and easily composable
pieces, where each component has a well defined abstraction. Pipelines are essentially configuration
pieces, where each component has a well-defined abstraction. Pipelines are essentially configuration
files specifying which Components should be used and when. This makes it easy to understand how your
code will execute while keeping it organized and structured in a maintainable way.
@ -32,7 +32,7 @@ Pipeline may run multiple Candidate Pipelines to fetch candidates to include in
Mixer Pipelines combine the results of multiple heterogeneous Candidate Pipelines together
(e.g. ads, tweets, users) while Recommendation Pipelines are used to score (via Scoring Pipelines)
and rank the results of homogenous Candidate Pipelines so that the top ranked ones can be returned.
and rank the results of homogenous Candidate Pipelines so that the top-ranked ones can be returned.
These pipelines also marshall candidates into a domain object and then into a transport object
to return to the caller.