Merge pull request #6 from kernc/requirements-help

Requirements help

README.md

@@ -1,29 +1,24 @@
 Orange3-Spark
-==================
+=============
 
-A set of widgets for Python's Orange Machine Learning to work over Apache Spark ML (Machine Learnint) API
+A set of widgets for Orange data mining suite to work with Apache Spark ML API.
 
 
+Requirements
+------------
 
-Requirements:
------------------
-
- - Python >=3.4
-
+ - Python >= 3.4
  - Pandas
-
- - Orange3
-
-Please follow the instruction to install Orange3 first.
-The main Orange proyect is hosted at: https://github.com/biolab/orange3
-
-
+ - Orange 3
 
+Please follow the instruction to install Orange 3 first.
 
+The main Orange project is hosted at: https://github.com/biolab/orange3
+Download from: http://orange.biolab.si
 
 
-It includes:
--------------------
+Features
+--------
 
   * A Spark Context.
   * A Hive Table.
@@ -39,12 +34,23 @@ It includes:
 
 ... more coming soon!
 
-Installing:
----------------
 
-pip install Orange3-spark
+Installing
+----------
+
+First, you need to have Apache Spark installed. Follow the instructions here:
+http://spark.apache.org/docs/latest/
+
+Then you can do:
+
+    pip install Orange3-spark
 
-or from the Orange3 add-on menu.
+or install the add-on from the Orange's Options | Add-ons menu. Note, if
+installing from Add-ons menu, the installation may fail if not all requirements
+are satisfiable.
 
-If install is Ok, you will see a new section in Orange ML containing a series of widgets from Spark ML API.
+If you require ODBC connectivity, you need to install `pyodbc`
+(which requires `sql.h` available if built with `pip` –
+that's `unixodbc-dev` package on Linux).
 
+If install is ok, you should see a new section in Orange containing a series of widgets from Spark ML API.

setup.py

@@ -63,7 +63,16 @@
             'orangecontrib.spark': ['tutorials/*.ows'],
             'orangecontrib.spark.widgets': ['icons/*'],
         },
-        install_requires = ['Orange', 'pandas', 'sqlparse', 'pyodbc'],
+        install_requires = [
+            'Orange',
+            'pandas',
+            'py4j',
+            'sqlparse',
+            'pyodbc',
+        ],
+        extras_require = {
+            'pyspark': [],
+        },
         entry_points = ENTRY_POINTS,
         keywords = ", ".join(KEYWORDS),
         namespace_packages = ['orangecontrib'],