Ran code_maintenance

git-svn-id: https://pymc.googlecode.com/svn/trunk@1281 15d7aa0b-6f1a-0410-991a-d59f85d14984

builddpkg

@@ -2,4 +2,4 @@
 
 python setupegg.py sdist_dsc --extra-cfg-file stdeb_all.cfg
 cd deb_dist/pymc-2.0
-dpkg-buildpackage -rfakeroot -uc -us
+dpkg-buildpackage -rfakeroot -uc -us

docs/INSTALL.tex

@@ -19,36 +19,36 @@ \section{Dependencies}
 Fortunately, there are currently only a few dependencies, and all are
 freely available online.
 \begin{itemize}
-\item {} 
+\item {}
 \href{http://www.python.org/.}{Python} version 2.5.
 
-\item {} 
+\item {}
 \href{http://www.scipy.org/NumPy}{NumPy} (1.2 or newer): The fundamental scientific programming package, it provides a
 multidimensional array type and many useful functions for numerical analysis.
 
-\item {} 
+\item {}
 \href{http://matplotlib.sourceforge.net/}{Matplotlib} (optional): 2D plotting library which produces publication
 quality figures in a variety of image formats and interactive environments
 
-\item {} 
+\item {}
 \href{http://www.pytables.org/moin}{pyTables} (optional): Package for managing hierarchical datasets and
 designed to efficiently and easily cope with extremely large amounts of data.
 Requires the \href{http://www.hdfgroup.org/HDF5/}{HDF5} library.
 
-\item {} 
+\item {}
 \href{http://code.google.com/p/pydot/}{pydot} (optional): Python interface to Graphviz's Dot language, it allows
 PyMC to create both directed and non-directed graphical representations of models.
 Requires the \href{http://www.graphviz.org/}{Graphviz} library.
 
-\item {} 
+\item {}
 \href{http://www.scipy.org/}{SciPy} (optional): Library of algorithms for mathematics, science
 and engineering.
 
-\item {} 
+\item {}
 \href{http://ipython.scipy.org/}{IPython} (optional): An enhanced interactive Python shell and an
 architecture for interactive parallel computing.
 
-\item {} 
+\item {}
 \href{http://somethingaboutorange.com/mrl/projects/nose/}{nose} (optional): A test discovery-based unittest extension (required
 to run the test suite).
 
@@ -99,10 +99,10 @@ \section{Installing from pre-built binaries}
 \usecounter{listcnt0}
 \setlength{\rightmargin}{\leftmargin}
 }
-\item {} 
+\item {}
 Download the installer for your platform from \href{http://pypi.python.org/pypi/pymc/}{PyPI}.
 
-\item {} 
+\item {}
 Double-click the executable installation package, then follow the
 on-screen instructions.
 

docs/README.tex

@@ -23,34 +23,34 @@ \section{Purpose}
 \section{Features}
 \label{features}
 \begin{itemize}
-\item {} 
+\item {}
 Fits Bayesian statistical models you create with Markov chain Monte Carlo and
 other algorithms.
 
-\item {} 
+\item {}
 Large suite of well-documented statistical distributions.
 
-\item {} 
+\item {}
 Gaussian processes.
 
-\item {} 
+\item {}
 Sampling loops can be paused and tuned manually, or saved and restarted later.
 
-\item {} 
+\item {}
 Creates summaries including tables and plots.
 
-\item {} 
+\item {}
 Traces can be saved to the disk as plain text, Python pickles, SQLite or MySQL
 database, or hdf5 archives.
 
-\item {} 
+\item {}
 Convergence diagnostics.
 
-\item {} 
+\item {}
 Extensible: easily incorporates custom step methods and unusual probability
 distributions.
 
-\item {} 
+\item {}
 MCMC loops can be embedded in larger programs, and results can be analyzed
 with the full power of Python.
 
@@ -63,20 +63,20 @@ \section{Features}
 \section{What's new in 2.0}
 \label{what-s-new-in-2-0}
 \begin{itemize}
-\item {} 
+\item {}
 New flexible object model and syntax (not backward-compatible).
 
-\item {} 
+\item {}
 Reduced redundant computations: only relevant log-probability terms are
 computed, and these are cached.
 
-\item {} 
+\item {}
 Optimized probability distributions.
 
-\item {} 
+\item {}
 New adaptive blocked Metropolis step method.
 
-\item {} 
+\item {}
 Much more!
 
 \end{itemize}

docs/database.tex

@@ -245,22 +245,22 @@ \subsection{mysql}
 port of the host server is given, along with a valid user name and password.
 These parameters are passed when the \texttt{Sampler} is instantiated:
 \begin{itemize}
-\item {} 
+\item {}
 \texttt{dbname} (\titlereference{string}) The name of the database file.
 
-\item {} 
+\item {}
 \texttt{dbuser} (\titlereference{string}) The database user name.
 
-\item {} 
+\item {}
 \texttt{dbpass} (\titlereference{string}) The user password for this database.
 
-\item {} 
+\item {}
 \texttt{dbhost} (\titlereference{string}) The location of the database host.
 
-\item {} 
+\item {}
 \texttt{dbport} (\titlereference{int})    The port number to use to reach the database host.
 
-\item {} 
+\item {}
 \texttt{dbmode} {\{}\texttt{a}, \texttt{w}{\}} File mode.  Use \texttt{a} to append values, and \texttt{w}
 to overwrite an existing database.
 
@@ -311,17 +311,17 @@ \subsection{hdf5}
 
 The \texttt{hdf5} Database takes the following parameters:
 \begin{itemize}
-\item {} 
+\item {}
 \texttt{dbname} (\titlereference{string}) Name of the hdf5 file.
 
-\item {} 
+\item {}
 \texttt{dbmode} {\{}\texttt{a}, \texttt{w}, \texttt{r}{\}} File mode: \texttt{a}: append, \texttt{w}: overwrite,
 \texttt{r}: read-only.
 
-\item {} 
+\item {}
 \texttt{dbcomplevel} : (\titlereference{int} (0-9)) Compression level, 0: no compression.
 
-\item {} 
+\item {}
 \texttt{dbcomplib} (\titlereference{string}) Compression library (\texttt{zlib}, \texttt{bzip2}, \texttt{lzo})
 
 \end{itemize}

docs/extending.tex

@@ -131,7 +131,7 @@ \subsection{General step methods} \label{user-gen}
 \begin{description}
    \item[\code{_id}:] A string that can identify each step method uniquely (usually something like \code{<class_name>_<stochastic_name>}).
    \item[\code{adaptive_scale_factor}:] An `adaptive scale factor'. This attribute is only needed if the default \code{tune()} method is used.
-   \item[\code{_tuning_info}:] A list of strings giving the names of any tuning parameters. For \texttt{Metropolis} instances, this would be \texttt{adaptive_scale_factor}. This list is used to keep traces of tuning parameters in order to verify `diminishing tuning' \cite{tuning}. 
+   \item[\code{_tuning_info}:] A list of strings giving the names of any tuning parameters. For \texttt{Metropolis} instances, this would be \texttt{adaptive_scale_factor}. This list is used to keep traces of tuning parameters in order to verify `diminishing tuning' \cite{tuning}.
 \end{description}
 
 All step methods have a property called \code{loglike}, which returns the sum of the log-probabilities of the union of the extended children of \code{self.stochastics}. This quantity is one term in the log of the Metropolis-Hastings acceptance ratio.

docs/modelfitting.tex

@@ -149,7 +149,7 @@ \section{Normal approximations} \label{sec:norm-approx}
 
 \code{MCMC} provides the following useful methods:
 \begin{description}
-    \item[\code{sample(iter, burn=0, thin=1, tune\_interval=1000, tune\_throughout=True, save\_interval=None, verbose=0)}:] Runs the MCMC algorithm and produces the traces. The \code{iter} argument controls the total number of MCMC iterations. No tallying will be done during the first \code{burn} iterations; these samples will be forgotten. After this burn-in period, tallying will be done each \code{thin} iterations. Tuning will be done each \code{tune\_interval} iterations. If \code{tune\_throughout=False}, no more tuning will be done after the burnin period. The model state will be saved every \code{save\_interval} iterations, if given.  
+    \item[\code{sample(iter, burn=0, thin=1, tune\_interval=1000, tune\_throughout=True, save\_interval=None, verbose=0)}:] Runs the MCMC algorithm and produces the traces. The \code{iter} argument controls the total number of MCMC iterations. No tallying will be done during the first \code{burn} iterations; these samples will be forgotten. After this burn-in period, tallying will be done each \code{thin} iterations. Tuning will be done each \code{tune\_interval} iterations. If \code{tune\_throughout=False}, no more tuning will be done after the burnin period. The model state will be saved every \code{save\_interval} iterations, if given.
     \item[\code{isample(iter, burn=0, thin=1, tune\_interval=1000, tune\_throughout=True, save\_interval=None, verbose=0)}:] An interactive version of \code{sample}. The sampling loop may be paused at any time, returning control to the user.
     \item[\code{use_step_method(method, *args, **kwargs)}:] Creates an instance of step method class \code{method} to handle some stochastic variables. The extra arguments are passed to the \code{init} method of \code{method}. Assigning a step method to a variable manually will prevent the \code{MCMC} instance from automatically assigning one. However, you may handle a variable with multiple step methods.
     % \item[\code{assign_step_methods()}:] Assigns step methods to all stochastic variables that do not currently have any. This method is called whenever \code{sample} or \code{isample} is called, but it can be useful to call it directly to see what the default step methods will be.
@@ -206,7 +206,7 @@ \subsection{Metropolis step methods} \label{metropolis}
    \item[\code{verbose}:] An integer.
 \end{description}
 
-Metropolis step methods adjust their initial proposal standard deviations using an attribute called \code{adaptive_scale_factor}. When \code{tune()} is called, the acceptance ratio of the step method is examined and this scale factor is updated accordingly. If the proposal distribution is normal, proposals will have standard deviation \code{self.proposal\_sd * self.adaptive_scale_factor}. 
+Metropolis step methods adjust their initial proposal standard deviations using an attribute called \code{adaptive_scale_factor}. When \code{tune()} is called, the acceptance ratio of the step method is examined and this scale factor is updated accordingly. If the proposal distribution is normal, proposals will have standard deviation \code{self.proposal\_sd * self.adaptive_scale_factor}.
 
 By default, tuning will continue throughout the sampling loop, even after the burnin period is over. This can be changed via the \texttt{tune\_throughout} argument to \texttt{MCMC.sample}. If an adaptive step method's \texttt{tally} flag is set (the default for \texttt{Metropolis}), a trace of its tuning parameters will be kept. If you allow tuning to continue throughout the sampling loop, it is important to verify that the `Diminishing Tuning' condition of \cite{tuning} is satisfied: the amount of tuning should decrease to zero, or tuning should become very infrequent.
 

pymc/CircularStochastic.py

@@ -11,48 +11,48 @@
 class CircularStochastic(Stochastic):
     """
     C = CircularStochastic(lo, hi, *args, **kwargs)
-    
+
     Takes two special parents, lo and hi; any incoming value
     will be mapped into the interval [lo,hi).
-    
+
     args and kwargs will be passed to Stochastic.__init__.
-    
+
     :SeeAlso: Stochastic
     """
     def __init__(self, lo, hi, *args, **kwargs):
         self.interval_parents = Container([hi, lo])
         Stochastic.__init__(self, *args, **kwargs)
-        
+
     def set_value(self, value):
         modded_value = mod_to_circle(value, *self.interval_parents.value).reshape(np.shape(value))
         Stochastic.set_value(self, modded_value)
     value = property(Stochastic.get_value, set_value)
-        
+
 class CircVonMises(CircularStochastic):
     """
     V = CircVonMises(name, mu, kappa, value=None, observed=False, size=1, trace=True, rseed=True, doc=None, verbose=0)
 
     Stochastic variable with Von Mises distribution.
     Parents are: mu, kappa.
-    
+
     Docstring of log-probability function:
     """+von_mises_like.__doc__
     def __init__(   self,
                     name,
                     mu, kappa,
                     value=None,
                     observed=False,
-                    size=1,                    
+                    size=1,
                     trace=True,
                     cache_depth=2,
                     rseed=True,
                     plot=None,
                     verbose = 0):
-                    
+
         if value is None:
             arg_eval = Container([mu, kappa]).value
             value = rvon_mises(arg_eval[0], arg_eval[1], size=size)
-    
+
         parents = {'mu':mu, 'kappa':kappa}
         logp = valuewrapper(von_mises_like)
         random = lambda mu, kappa, size=size: rvon_mises(mu, kappa, size=size)

pymc/CommonDeterministics.py

@@ -434,7 +434,7 @@ def wrapper(**kwargs):
 %s
     """%(func.__name__, '_'*60, func.__doc__)
     return dtrm_generator
-        
+
 
 def pfunc(func):
     """
@@ -468,7 +468,7 @@ def pfunc(func):
             raise cls, inst, tb
     fargs, fvarargs, fvarkw, fdefaults = inspect.getargspec(func)
     n_fargs = len(fargs)
-    
+
     def dtrm_generator(*args, **kwds):
         name = func.__name__ + '('+'_'.join([str(arg) for arg in list(args) + kwds.values()])+')'
         doc_str = 'A deterministic returning %s(%s, %s)'%(func.__name__, ', '.join([str(arg) for arg in args]), ', '.join(['%s=%s'%(key, str(val)) for key, val in kwds.iteritems()]))

pymc/Container_values.pyx

@@ -18,7 +18,7 @@ cdef class LCValue:
     cdef int n_val, n_nonval
 
     def __init__(self, container):
-        
+
         cdef int i
         self._value = container._value
         self.val_ind = <int*> PyArray_DATA(container.val_ind)
@@ -27,7 +27,7 @@ cdef class LCValue:
         self.nonval_ind = <int*> PyArray_DATA(container.nonval_ind)
         self.nonval_obj = <void**> PyArray_DATA(container.nonval_obj)
         self.n_nonval = container.n_nonval
-                    
+
     def run(self):
         cdef int i
         for i from 0 <= i < self.n_val:

pymc/MCMC.py

@@ -64,7 +64,7 @@ def __init__(self, input=None, db='ram', name='MCMC', calc_deviance=True, **kwds
               Keywords arguments to be passed to the database instantiation method.
         """
         Sampler.__init__(self, input, db, name, calc_deviance=calc_deviance, **kwds)
-        
+
         self._sm_assigned = False
         self.step_method_dict = {}
         for s in self.stochastics:
@@ -100,7 +100,7 @@ def assign_step_methods(self):
         Make sure every stochastic variable has a step method. If not,
         assign a step method from the registry.
         """
-        
+
         if not self._sm_assigned:
 
             # Assign dataless stepper first
@@ -132,7 +132,7 @@ def assign_step_methods(self):
             self.step_methods = set()
             for s in self.stochastics:
                 self.step_methods |= set(self.step_method_dict[s])
-            
+
             for sm in self.step_methods:
                 if sm.tally:
                     for name in sm._tuning_info:
@@ -146,7 +146,7 @@ def sample(self, iter, burn=0, thin=1, tune_interval=1000, tune_throughout=True,
         sample(iter, burn, thin, tune_interval, tune_throughout, save_interval, verbose)
 
         Initialize traces, run sampling loop, clean up afterward. Calls _loop.
-        
+
         :Parameters:
           - iter : int
             Total number of iterations to do
@@ -157,7 +157,7 @@ def sample(self, iter, burn=0, thin=1, tune_interval=1000, tune_throughout=True,
           - tune_interval : int
             Step methods will be tuned at intervals of this many iterations, default 1000
           - tune_throughout : boolean
-            If true, tuning will continue after the burnin period (True); otherwise tuning 
+            If true, tuning will continue after the burnin period (True); otherwise tuning
             will halt at the end of the burnin period.
           - save_interval : int or None
             If given, the model state will be saved at intervals of this many iterations
@@ -201,8 +201,8 @@ def _loop(self):
                 # Tune at interval
                 if i and not (i % self._tune_interval) and self._tuning:
                     self.tune()
-                
-                if i == self._burn: 
+
+                if i == self._burn:
                     if self.verbose>0:
                         print 'Burn-in interval complete'
                     if not self._tune_throughout:

pymc/Matplot.py

@@ -812,7 +812,7 @@ def zplot(pvalue_dict, name='', format='png', path='./', fontmap = {1:10, 2:8, 3
         vname = var
         vname += " (%i)" % size(zvals)
         labels = append(labels, vname)
-        
+
     # Spawn new figure
     figure()
     subplot(111)

pymc/Model.py

@@ -378,7 +378,7 @@ def _assign_database_backend(self, db):
 
         # Assign Trace instances to tallyable objects.
         self.db.connect_model(self)
-        
+
 
     def pause(self):
         """Pause the sampler. Sampling can be resumed by calling `icontinue`.
@@ -619,7 +619,7 @@ def trace(self, name, chain=-1):
             return self.db.trace(name.__name__, chain)
         else:
             raise ValueError, 'Name argument must be string or Variable, got %s.'%name
-            
+
     def _get_deviance(self):
         return self._sum_deviance()
     deviance = property(_get_deviance)