Merge pull request #23 from jchodera/fix-conda-recipe

Fix conda recipe

bhmm/__init__.py

@@ -30,4 +30,4 @@
 
 # other stuff
 from bhmm.util import config
-from bhmm.util import testsystems
+from bhmm.util import testsystems

bhmm/api.py

@@ -2,12 +2,9 @@
 
 import numpy as _np
 
-from bhmm.util import types as _types
-from bhmm.hmm.generic_hmm import HMM as _HMM
-from bhmm.estimators.maximum_likelihood import MaximumLikelihoodEstimator as _MaximumLikelihoodEstimator
-from bhmm.estimators.bayesian_sampling import BayesianHMMSampler as _BHMM
-
 def _guess_model_type(observations):
+    from bhmm.util import types as _types
+
     o1 = _np.array(observations[0])
 
     # CASE: vector of int? Then we want a discrete HMM
@@ -109,6 +106,7 @@ def gaussian_hmm(P, means, sigmas, pi=None, stationary=True, reversible=True):
     # initialize output model
     output_model = GaussianOutputModel(nstates, means, sigmas)
     # initialize general HMM
+    from bhmm.hmm.generic_hmm import HMM as _HMM
     ghmm = _HMM(P, output_model, Pi=pi, stationary=stationary, reversible=reversible)
     # turn it into a Gaussian HMM
     ghmm = GaussianHMM(ghmm)
@@ -136,6 +134,7 @@ def discrete_hmm(P, pout, pi=None, stationary=True, reversible=True):
     # initialize output model
     output_model = DiscreteOutputModel(pout)
     # initialize general HMM
+    from bhmm.hmm.generic_hmm import HMM as _HMM
     dhmm = _HMM(P, output_model, Pi=pi, stationary=stationary, reversible=reversible)
     # turn it into a Gaussian HMM
     dhmm = DiscreteHMM(dhmm)
@@ -191,6 +190,7 @@ def estimate_hmm(observations, nstates, lag=1, initial_model=None, type=None,
         observations = _lag_observations(observations, lag)
 
     # construct estimator
+    from bhmm.estimators.maximum_likelihood import MaximumLikelihoodEstimator as _MaximumLikelihoodEstimator
     est = _MaximumLikelihoodEstimator(observations, nstates, initial_model=initial_model, type=type,
                                       reversible=reversible, stationary=stationary, p=p, accuracy=accuracy, maxit=maxit)
     # run
@@ -222,6 +222,7 @@ def bayesian_hmm(observations, estimated_hmm, nsample=100, store_hidden=False):
 
     """
     # construct estimator
+    from bhmm.estimators.bayesian_sampling import BayesianHMMSampler as _BHMM
     sampler = _BHMM(observations, estimated_hmm.nstates, initial_model=estimated_hmm,
                     reversible=estimated_hmm.is_reversible, transition_matrix_sampling_steps=1000,
                     type=estimated_hmm.output_model.model_type)

devtools/conda-recipe/meta.yaml

@@ -9,29 +9,24 @@ build:
 requirements:
   build:
     - python
+    - setuptools
     - cython
     - numpy
     - scipy
-    - setuptools
-    - docopt
-    - jinja2
+    - pyemma >=1.2
     - scikit-learn
-    - matplotlib
-    - seaborn
-    - pyemma >=1.1.2
+    # - matplotlib
+    # - seaborn
 
   run:
     - python
+    - cython
     - numpy
     - scipy
-    - cython
-    - docopt
-    - pyyaml
-    - jinja2
+    - pyemma >=1.2
     - scikit-learn
-    - matplotlib
-    - seaborn
-    - pyemma >=1.1.2
+    # - matplotlib
+    # - seaborn
 
 test:
   requires:
@@ -44,7 +39,4 @@ test:
 about:
   home: https://github.com/choderalab/bhmm
   license: GNU Lesser General Public License v3 or later (LGPLv3+)
-
-
-
-
+  summary: Bayesian hidden Markov models for analysis of single-molecule trajectory data

setup.py

@@ -16,7 +16,7 @@
 DOCLINES = __doc__.split("\n")
 
 ########################
-VERSION = "0.1.0"
+VERSION = "0.2.0"
 ISRELEASED = False
 __version__ = VERSION
 ########################
@@ -136,6 +136,7 @@ def find_package_data(data_root, package_root):
                         include_dirs = [numpy.get_include()])]
 
 write_version_py()
+
 setup(
     name='bhmm',
     author='John Chodera and Frank Noe',
@@ -149,18 +150,16 @@ def find_package_data(data_root, package_root):
     classifiers=CLASSIFIERS.splitlines(),
     package_dir={'bhmm': 'bhmm'},
     #packages=['bhmm', "bhmm.tests"] + ['bhmm.%s' % package for package in find_packages('bhmm')],
-    packages=['bhmm', 'bhmm.tests', 'bhmm.msm', 'bhmm.hidden', 'bhmm.init', 'bhmm.msm', 'bhmm.output_models', 'bhmm.output_models.impl_c', 'bhmm.util', 'bhmm.hidden.impl_python', 'bhmm.hidden.impl_c'],
+    packages=['bhmm', 'bhmm.tests', 'bhmm.hmm', 'bhmm.estimators', 'bhmm.msm', 'bhmm.hidden', 'bhmm.init', 'bhmm.msm', 'bhmm.output_models', 'bhmm.output_models.impl_c', 'bhmm.util', 'bhmm.hidden.impl_python', 'bhmm.hidden.impl_c'],
     # + ['bhmm.%s' % package for package in find_packages('bhmm')],
-    package_data={'bhmm': find_package_data('examples', 'bhmm')},  # NOTE: examples installs to bhmm.egg/examples/, NOT bhmm.egg/bhmm/examples/.  You need to do utils.get_data_filename("../examples/*/setup/").
+    package_data={'bhmm': find_package_data('examples', 'bhmm') + find_package_data('bhmm/tests/data', 'bhmm')},  # NOTE: examples installs to bhmm.egg/examples/, NOT bhmm.egg/bhmm/examples/.  You need to do utils.get_data_filename("../examples/*/setup/").
     zip_safe=False,
     install_requires=[
         'cython',
         'numpy',
         'scipy',
         'pyemma>=1.2',
         'scikit-learn',
-        'nose',
-        'docopt>=0.6.1',
         ],
     ext_modules = cythonize(extensions)
     )