Add a README

diff --git a/src/ShapeRecogn/README.md b/src/ShapeRecogn/README.md
new file mode 100644 (file)
index 0000000..b885d22
--- /dev/null
+++ b/src/ShapeRecogn/README.md
@@ -0,0 +1,99 @@
+# ShapeRecogn
+A tool leveraging the MEDCoupling library for recognizing canonical shapes in 3D mesh files.
+
+## Table Of Contents
+
+1. [Quickstart](#quickstart)
+2. [Installation](#installation)
+3. [Description](#description)
+
+## Quickstart
+
+### With output file
+
+```python
+>> import ShapeRecogn as sr
+>>
+>> shape_recogn = sr.ShapeRecognMesh("resources/ShapeRecognCone.med")
+>> shape_recogn.recognize("ShapeRecognCone_areas.med")
+```
+
+### Without output file
+
+```python
+>> import ShapeRecogn as sr
+>>
+>> shape_recogn = sr.ShapeRecognMesh("resources/ShapeRecognCone.med")
+>> shape_recogn.recognize()
+>> radius_field = shape_recogn.buildRadius()
+>> radius_field
+MEDCouplingFieldDouble C++ instance at 0x55f3418c6700. Name : "Radius (Area)".
+Nature of field : NoNature.
+P1 spatial discretization.
+
+Mesh info : MEDCouplingUMesh C++ instance at 0x55f34171a2d0. Name : "Mesh_1". Mesh dimension : 2. Space dimension : 3.
+
+Array info : DataArrayDouble C++ instance at 0x55f3418b40e0. Number of tuples : 957. Number of components : 1.
+[9.8948383131651862, 9.8948383131651862, 9.8948383131651862, 9.8948383131651862, 9.8948383131651862, 9.8948383131651862, 9.8948383131651862, 9.8948383131651862, 9.8948383131651862, 9.8948383131651862, 9.8948383131651862, 9.8948383131651862, 9.8948383131651862, 9.8948383131651862, ... ]
+```
+
+## Installation
+
+The installation 
+
+## Description
+
+ShapeRecogn is a tool that leverages the MEDCoupling library to recognize
+canonical shapes from 3D meshes using triangle cells.
+
+The tool is based on the thesis work of Roseline Bénière *[Reconstruction d’un modèle B-Rep à partir d’un maillage 3D](https://theses.fr/2012MON20034)*, and it recognizes five canonical shapes:
+ - Plane (0)
+ - Sphere (1)
+ - Cylinder (2)
+ - Cone (3)
+ - Torus (4)
+ - Unknown (5) (When the algorithm failed the recognition)
+
+The recognized shapes are divided into areas within the mesh.
+The tool also generates an output file with detailed fields describing the areas and their properties. This makes it easier to analyze or further process the mesh data.
+ - Area Id: To distinguish the areas
+ - Primitive Type (Area): One of the canonical shape with the id describe above
+ - Normal (Area):
+    * Normal of a plane area
+ - Minor Radius (Area)
+    * Minor radius of a torus area
+ - Radius (Area)
+    * Radius of a sphere area
+    * Radius of a cylinder area
+    * Radius of the base of a cone area
+    * Major radius of a torus area
+ - Angle (Area)
+    * Angle between the central axis and the slant of a cone area
+ - Center (Area)
+    * Center of a sphere area
+    * Center of a torus area
+ - Axis (Area)
+    * Central axis of a cylinder area
+    * Central axis of a cone area
+ - Apex (Area)
+    * Apex of a cone area
+
+Some intermediate computation values concerning the nodes are also available:
+ - K1 (Node) and K2 (Node): the *[principal curvatures](https://en.wikipedia.org/wiki/Principal_curvature)*
+ - Primitive Type (Node) : One of the canonical shape with the id describe above. The primitive type is deduced usint K1 and K2 values.
+ - Normal (Node): Normal of the nodes using neighbor nodes
+
+Each field can be respectively build as a `MEDCouplingDoubleField` using the following methods of the `ShapeRecognMesh` class:
+ - `buildAreaId()`
+ - `buildAreaPrimitiveType()`
+ - `buildAreaNormal()`
+ - `buildMinorRadius()`
+ - `buildRadius()`
+ - `buildAngle()`
+ - `buildCenter()`
+ - `buildAxis()`
+ - `buildApex()`
+ - `buildNodeK1()`
+ - `buildNodeK2()`
+ - `buildNodePrimitiveType()`
+ - `buildNodeNormal()`