Fili
Read this in other languages: 中文.
The Fili groupId has been changed from com.yahoo.fili to com.yahoo.bard.
Fili is a Java-based framework that makes it easy to build and maintain RESTful web services for time series reporting and analytics. Fili's HTTP GET-based reporting API is clean and user-friendly, hiding the complexities of complicated metric and dimension definition, data storage, and query optimization from end-users. Designed with Big Data and scalability in mind, Fili has first-class support for Druid as a back-end, but Fili's flexible pipeline-style architecture can handle nearly any back-end for data storage.
Fili exposes the same core concepts that all time series reporting and analytics systems expose:
- Metrics
- Dimensions
- Tables
- Time (Reporting Time Grain and Interval)
Other systems expose additional concepts like Views, Partitions, and metric formulas, but Fili chooses not to expose end-users to those low-level concerns. By limiting the mental model of Fili's API to just these core domain concepts, Fili's API allows end-users to better extract business value by focusing on finding answers to the what and why questions from their data, rather than on how to ask their question.
Fili's simple and clear API lets users focus on their business-driven questions, and the Fili library takes care of figuring out how to best answer the question for them.
This simplicity also allows for a huge amount of flexibility around where data is stored and how it's queried and retrieved from back-end systems. Having such flexibility allows those maintaining a web service built with Fili to move data around, better optimize queries, and even swap out entire back-end systems without breaking their end-users or forcing them to go through a painful migration.
Fili also provides a plethora of other capabilities. Here are some of them:
| Functionality | Operability |
|---|---|
| Complex metric definition | Rate limiting |
| Performance slice routing | Query weight checks |
| Dimension joins (both annotation and filtering) | Rich usage metrics |
| Partial interval protection | Health checks |
| Volatile data handling | Caching |
| Modular architecture |
Community 
The Fili community generally hangs out on Gitter, so please drop by if you have any questions, comments, concerns, wishes, hopes, dreams, wants, needs, yearnings, musings, or idle curiosities about Fili. The core Fili developers love meeting new people and talking with them about how they can best use Fili to solve their Big Data time series reporting and analytics problems. We know there are gaps in Fili, and definitely lots of new, powerful capabilities to add, so hearing about what's working and what could be better will help drive the direction of Fili.
If you have other thoughts, or are running into trouble and are not able to get help from the community on Gitter, please open an issue describing your problem or idea.
If you would like to get involved with Fili development, check out the CONTRIBUTING file.
Quick Start
Fili comes with a pre-configured example application to help you get started and serve as a jumping-off-point for building your own web service using Fili. The example application lets you report on Wikipedia article edit information, and picks up where Druid's quick-start tutorial leaves off.
Versioning
Fili is stable and production-ready today, but the codebase is still under active development, with many large-scale changes and new features being worked on or planned for the future.
Active development happens on the patch version of the highest minor version.
@Deprecated
APIs marked with the @Deprecated annotation are planned for removal in upcoming releases. Deprecated APIs will be
supported for 1 stable release beyond the release in which they were deprecated, but it is strongly recommended to stop
using them. After deprecated APIs are no longer supported, they may be removed at any time.
Binaries (How to Get It)
Binaries for Fili are stored in Maven Central. Dependency information for Maven, Ivy, and Gradle can be found at https://search.maven.org/search?q=com.yahoo.bard, and some examples are below.
Maven:
<dependency>
<groupId>com.yahoo.bard</groupId>
<artifactId>fili</artifactId>
<version>x.y.z</version>
</dependency>
<repository>
<id>fili</id>
<url>https://repo1.maven.org/maven2</url>
</repository>Gradle:
repositories {
maven { url 'https://repo1.maven.org/maven2' }
}
dependencies {
compile 'com.yahoo.bard:fili:x.y.z'
}Extending
Fili's easy to extend! It has a bunch of hooks already (AbstractBinderFactory)!
There's a module system for config and modules depending on other modules! There's lots more to say here, but the time
the time, look at the time!
Documentation
More information about Fili can be found at fili.io.
Contributing
More details can be found in the CONTRIBUTING file.
LICENSE
Copyright 2016 Yahoo! Inc.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.