Reo is a programming language and environment for data analysis.
- Data import/export - databases, Excel/CSV/JSON/XML files, HTTP APIs etc.
- A simple and expressive functional language for data analysis.
- High-performance immutable data structures - lists, vectors, sets, dictionaries, tables.
- Built-in statistical functions.
- Charting capabilities.
- Parallel and distributed computing.
- Seamless integration with the Java ecosystem and external tools.
Make sure you have JDK 10 or later installed. You can build and run Reo with the following commands:
$ make
$ ./reo
This will land you in the REPL, where you can play with the platform:
reo v0.0.1
=>
The => prompt shows that Reo is waiting for your input.
It is suggested that you work through the following short tutorial and then proceed to the rest of the Documentation.
In this short tutorial, let's create and manipulate a database for managing employee salaries. A columnar-table is one of the basic structures used by Reo. This format makes it easy and efficient to write functions that aggregate values in individual columns.
We use the tab function to manually create a table:
=> emp:tab(['name 'salary]
[["Max G" "Kevin J" "Sue D" "Ben B" "Joe G"]
[1900.78 2344.88 1200.56 3400.56 1300.78]])The first argument is a vector (or a collection) of column names. We have two symbolic names here: 'name and 'salary.
The second argument is a vector of vectors. Each vector is the values for each column.
Let's have a look at the table now:
=> emp
; name: [Max G Kevin J Sue D Ben B Joe G]
; salary: [1900.78 2344.88 1200.56 3400.56 1300.78]Note In this tutorial, the output from Reo will be identified by a semi-colon (;). This is not part of the output produced by the real REPL.
Here is a quick "statistical" summary of the table:
=> summary(emp)
; [[col:name count:0 is_numeric:0b Joe_G:1 Ben_B:1 Sue_D:1 Kevin_J:1 Max_G:1]
; [col:salary min:1200.56 max:3400.56 mean:2029.512 median:1900.78 is_numeric:1b]]We may also get graphical views of our data, in spreadsheet format or as a chart:
=> view(emp)
=> chart('bar emp('name) emp('salary))
If you want to, you can just flip a table to a record based format.
=> flip(emp)
; name salary
; ---------------------------
; Max G 1900.78
; Kevin J 2344.88
; Sue D 1200.56
; Ben B 3400.56
; Joe G 1300.78
=> flip(flip(emp))
; name: [Max G Kevin J Sue D Ben B Joe G]
; salary: [1900.78 2344.88 1200.56 3400.56 1300.78]Record-based tables might be the ideal structure for some operations, but in this tutorial we only work with columnar data.
Let's write a function to compute a given percentage of each salary:
=> sals:emp('salary)
=> incr:fn(percent) map(fn(s) s * percent, sals)We can use this function to give a 10% salary increment to all employees:
=> new_sals:incr(0.1) + sals
=> new_sals
; [2090.858 2579.368 1320.616 3740.616 1430.858]How much salary increase is allotted to each employee?
=> new_sals - sals
; [190.078 234.48 120.05 340.05 130.077]What is the total additional cost incurred to the company by the salary increase?
=> sum(new_sals - sals)
; 1014.75That concludes our very short tutorial! To learn more, please go the detailed documentation.
The basic design of Reo is inspired by the APL family of array programming languages. Another major influence is Lisp, especially the Scheme and Clojure variants of that language.
Reo itself is implemented in Clojure and makes extensive use of its core functions, immutable datastructures and evaluator. Reo runs on top of the Java Virtual Machine and is integrated well into that high-quality platform.
Copyright (c) 2019 Anvetsu Technologies Pvt Ltd (https://anvetsu.com/).
Please see the LICENSE file for licensing terms.