Merge pull request #143 from pierrepo/remove_clust_R

Remove R clustering in main code, tests and doc

.travis.yml

@@ -31,7 +31,6 @@ matrix:
 addons:
     apt:
         packages:
-            - r-base
             - ghc
             - cabal-install
 
@@ -66,7 +65,6 @@ script:
     - if [[ $SETUP == 'test' ]]; then nosetests -v pbxplore/tests; fi
     - if [[ $SETUP == 'test' ]]; then yes | ./run_demo1_assignation.sh; fi
     - if [[ $SETUP == 'test' ]]; then yes | ./run_demo2_statistics.sh; fi
-    - if [[ $SETUP == 'test' ]]; then yes | ./run_demo2_clusters.sh; fi
     - if [[ $SETUP == 'doc' ]]; then cd doc; sphinx-build -W -b html source build/html; fi
 
 #after_success:

CHANGELOG

@@ -6,6 +6,7 @@
 - Add MDAnalysis as a mandatory module
 - Add matplotlib as a mandatory module
 - Fix weblogo weblogolib.SymbolColor/ColorGroup import 
+- Remove R clustering (now in branch clust_R)
 
 **1.3.1**
 - remove pbxplore import from setup.py and add version number

doc/source/api_cookbook.rst

@@ -30,13 +30,12 @@ Look at the notebook  :doc:`Writing PB in file <./notebooks/Assignement>` for fu
 PBxplore.analysis
 -----------------
 
-This module handle all analysis functions, ploting functions and clustering one available with the package.
+This module handle all analysis functions and ploting functions.
 You can:
 
 * generate map of the distribution of PBs along protein sequence with `plot_map()`.
 * compute :ref:`Neq` with `compute_neq()` and generate the plot with `plot_neq()`.
 * generate WebLogo-like representation of PBs frequency along protein sequence with `generate_weblogo()`.
-* cluster conformations of a same protein based on PB similarities.
 
 Look at the notebook :doc:`Visualize protein deformability <./notebooks/Deformability>`
 for further information.

doc/source/index.rst

@@ -47,7 +47,6 @@ Basically, PBxplore can:
    PBassign
    PBcount
    PBstat
-   PBclust
    api_cookbook
    api_reference
 

doc/source/utilization.rst

@@ -17,7 +17,6 @@ Here the list:
 - :doc:`PBcount <PBcount>` computes the frequency of PBs at each position
   along the amino acid sequence.
 - :doc:`PBstat <PBstat>` generates frequency and logo plots, and estimates the :ref:`Neq <Neq>`.
-- :doc:`PBclust <PBclust>` use clustering algorithm (k-means) to re-group similar PBs sequences.
 
 
 API
@@ -35,14 +34,12 @@ Demo files
 PBxplore provides scripts to demonstrate its functionalities. These scripts
 guide the user through the different command line tools of PBxplore.
 
-3 demonstration scripts are available:
+2 demonstration scripts are available:
 
 * `run_demo1_assignation.sh` demonstrates the how to use ``PBassign`` to assign
   Protein Block sequences to protein structures;
 * `run_demo2_statistics.sh` demonstrates how to analyse protein dynamics using
-  PBxplore;
-* finally, `run_demo2_clusters.sh` demonstrates how to use ``PBclust`` to cluster
-  Protein Block sequences.
+  PBxplore.
 
 In addition to the scripts, PBxplore bundles input files to test and get
 accustom with the software and the python library. Call the ``PBdata`` program to

pbxplore/analysis/__init__.py

@@ -28,12 +28,6 @@
 
 .. autofunction:: generate_weblogo
 
-Cluster protein block sequences
--------------------------------
-
-.. autofunction: distance_matrix
-
-.. autofunction: hclust
 
 Utils
 -----
@@ -45,7 +39,6 @@
 .. autofunction:: compute_score_by_position
 """
 
-from .clustering import hclust, distance_matrix, RError
 from .compare import compare
 from .count import count_matrix, read_occurence_file
 from .neq import compute_neq