Cannot Install HDDM in Windows

[Long Ding]

Hi, hddm-users,

Has anyone successfully installed hddm in Windows recently? 

I've tried many versions of Anaconda (Anaconda-2.1.0, 2.2.0, 2.3.0 and 2.4.0 Windows-x86). I used the following conda commands as suggested in Readme:

conda install pymc 

conda install -c pymc hddm

There was no error in building the package, but when I do "Import hddm" in Python, either Python hangs with the "In:" prompt or I get the following errors: 

Python 2.7.10 |Anaconda 2.4.0 (32-bit)| (default, Nov  7 2015, 13:29:20) [MSC v.

1500 32 bit (Intel)] on win32

Type "help", "copyright", "credits" or "license" for more information.

Anaconda is brought to you by Continuum Analytics.

Please check out: http://continuum.io/thanks and https://anaconda.org

>>> import hddm

Traceback (most recent call last):

 File "<stdin>", line 1, in <module>

 File "C:\Users\Long\Anaconda2\lib\site-packages\hddm\__init__.py", line 7, in

<module>

   import likelihoods

 File "C:\Users\Long\Anaconda2\lib\site-packages\hddm\likelihoods.py", line 2,

in <module>

   import pymc as pm

 File "C:\Users\Long\Anaconda2\lib\site-packages\pymc\__init__.py", line 33, in

<module>

   from .Model import *

 File "C:\Users\Long\Anaconda2\lib\site-packages\pymc\Model.py", line 17, in <m

odule>

   from . import database

 File "C:\Users\Long\Anaconda2\lib\site-packages\pymc\database\__init__.py", li

ne 56, in <module>

   from . import hdf5

 File "C:\Users\Long\Anaconda2\lib\site-packages\pymc\database\hdf5.py", line 2

5, in <module>

   import tables

 File "C:\Users\Long\Anaconda2\lib\site-packages\tables\__init__.py", line 84,

in <module>

   from tables.utilsextension import (

 File "__init__.pxd", line 155, in init tables.utilsextension (tables\utilsexte

nsion.c:17736)

ValueError: numpy.dtype has the wrong size, try recompiling

I saw from other threads that pymc 2.3.5 might have a problem. I tried 

conda install pymc==2.3.4 

conda install -c pymc hddm pymc=2.3.4

Still no go. 

Any suggestions? 

Thanks in advance!

Long

[Sam Mathias]

Please post your solution.

On 18 November 2015 at 15:10, Long Ding <longde...@gmail.com> wrote:

Never mind. I figured out a way. 

--
You received this message because you are subscribed to the Google Groups "hddm-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hddm-users+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Long Ding]

I figured out that I can install hddm if I just stick with python 2.7.3 and pymc 2.3.4. So far I can import hddm and load an example csv file. 

Long

On Tuesday, November 17, 2015 at 9:33:24 PM UTC-5, Long Ding wrote:

[Thomas Wiecki]

Thanks, Long. How did you force anaconda to python 2.7.3?

[Long Ding]

Sorry about the brief response before. I have access to my laptop now, here are the more detailed work-around:

The anaconda I used is: Anaconda2-2.4.0-Windows-x86.exe, which is a 32-bit Anaconda version 2.4.0 with Python 2.7.10.

In the Anaconda prompt, run

conda create --name py27_3 python=2.7.3

conda install -n py27_3 pymc==2.3.4 python=2.7.3

conda install -n py27_3 -c pymc hddm python=2.7.3 pymc=2.3.4

These steps install the hddm package in an environment called "py27_3". To use the hddm package within the environment:

activate py27_3

python

import hddm

Note that the path setting is different in this environment from what the hddm README assumes. So to run the sample codes on the example file simple_difficult.csv, specify the full path for the data file. 

data=hddm.load_csv('C:\Anaconda2\pkgs\hddm-0.5.5-np19py27_0\Lib\site-packages\hddm\examples\simple_difficulty.csv')

The other lines of sample codes can be run as is. 

I have not tried later versions of python 2.7. Just picked the earliest possible python 2.7 version in this version of Anaconda. 

As my criterion was just to get hddm installed without blatant error messages, I would be very interested to know from hddm experts if this work-around may compromise some hddm functions. 

Thanks!

Long

 

[Thomas Wiecki]

Thanks very much for the detailed description. That should help any windows user to get hddm to work again.

--

[Long Ding]

Hi, hddm experts,  

I just ran into another error. Was just testing with example data and code (see below). model.print_stats() returned correct values, but the geweke statistic computation resulted in errors. Could it be because of some file version issue? Thanks for any suggestions!

Long

Code: 

import hddm

data=hddm.load_csv('simple_difficulty.csv')

model = hddm.HDDM(data, depends_on={'v':'difficulty'})

model.find_starting_values()

model.sample(2000, burn=20)

model.print_stats()

from kabuki.analyze import check_geweke

print check_geweke(model)

Error message:

Traceback (most recent call last):

  File "<ipython-input-71-abf70ec3d99b>", line 2, in <module>

    print check_geweke(model)

  File "C:\Users\Long\Anaconda2\envs\py27_3\lib\site-packages\kabuki\analyze.py", line 160, in check_geweke

    if np.any(np.abs(geweke[:,1]) > 2):

IndexError: too many indices for array

statsmodels not available. Geweke diagnostic cannot be calculated.

 

[Sam Mathias]

Looks like you need to install statsmodels.

Samuel R. Mathias

Postdoctoral Associate

Department of Psychiatry, Yale School of Medicine

Suite 3014, 2 Church Street South

New Haven, CT 06519

http://www.srmathias.com

--

[Thomas Wiecki]

It seems like anaconda has an updated version of pymc 2.3.6. Can someone see if it works and let me know?

[Long Ding]

You are absolutely right! Thanks!

[stein...@gmail.com]

Hi together,

following your instructions I get this error. Any help?
Thanks a lot
- Benjamin

[Sam Mathias]

That is actually a depreciation warning, not an error. You can still use the python interactive shell just fine: try typing a command. 

[stein...@gmail.com]

right, now I just have to figure out why it does not seem to work with IPython notebook. When I choose thepy27_3 environment in the anaconda launcher a commandwindow just briefly pops up and vanishes again. -now browser tab opening - nothing left behind :-(

[Sam Mathias]

Sorry, I don't understand what you are trying to do. Can you post your code?

[stein...@gmail.com]

Hi Samuel thanks for taking the time.

On my mac I use HDDM within IPython notebook.

Trying to install it on this windows 7 PC I followed the instructions given by Long Ding above:

conda create --name py27_3 python=2.7.3

conda install -n py27_3 pymc==2.3.4 python=2.7.3

conda install -n py27_3 -c pymc hddm python=2.7.3 pymc=2.3.4

I also tried thies procedure with both Version 2.3.4 and version 2.3.6 of pymc.

When I open the Anaconda Launcher now to start an IPython notebook I can choose between the original environment ("root") where python 2.7.10 is installed and the new Environment ("py27_3") where python 2.7.3 is installed (see picture). If I choose the new Environment and launch an IPython notebook I get just a glimpse of a command window and nothing else. using the old environment still launches a notebook in my browser. But which leaves me with an error when trying to import kabuki (see further down below). I think that is the Problem Long Ding encountered as well.

Am Mittwoch, 25. November 2015 19:32:35 UTC+1 schrieb Samuel

---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
RuntimeError: module compiled against API version a but this version of numpy is 9

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-4-045d55581b72> in <module>()
----> 1 import kabuki

C:\Anaconda2\lib\site-packages\kabuki\__init__.py in <module>()
----> 1 from hierarchical import *
      2 
      3 import utils
      4 import analyze
      5 import step_methods as steps

C:\Anaconda2\lib\site-packages\kabuki\hierarchical.py in <module>()
     11 
     12 import pandas as pd
---> 13 import pymc as pm
     14 import warnings
     15 

C:\Anaconda2\lib\site-packages\pymc\__init__.pyc in <module>()
     31 from .NumpyDeterministics import *
     32 from .distributions import *
---> 33 from .Model import *
     34 from .StepMethods import *
     35 from .MCMC import *

C:\Anaconda2\lib\site-packages\pymc\Model.py in <module>()
     15 from numpy import zeros, floor
     16 from numpy.random import randint
---> 17 from . import database
     18 from .PyMCObjects import Stochastic, Deterministic, Node, Variable, Potential
     19 from .Container import Container, ObjectContainer

C:\Anaconda2\lib\site-packages\pymc\database\__init__.py in <module>()
     54 
     55 try:
---> 56     from . import hdf5
     57 except ImportError:
     58     pass

C:\Anaconda2\lib\site-packages\pymc\database\hdf5.py in <module>()
     23 from pymc.database import base, pickle
     24 from copy import copy
---> 25 import tables
     26 import os
     27 import warnings

C:\Anaconda2\lib\site-packages\tables\__init__.py in <module>()
     82 
     83 # Necessary imports to get versions stored on the cython extension
---> 84 from tables.utilsextension import (
     85     get_pytables_version, get_hdf5_version, blosc_compressor_list,
     86     blosc_compcode_to_compname_ as blosc_compcode_to_compname,

__init__.pxd in init tables.utilsextension (tables\utilsextension.c:17736)()

ValueError: numpy.dtype has the wrong size, try recompiling

Mathias:

[Thomas Wiecki]

I have seen this is a couple of times with anaconda. I don't it's directly related to HDDM but something that installed it's own numpy. Can you try conda install -f numpy

Thomas

[stein...@gmail.com]

Hi Thomas,

with python 2.7.10 the kabuki problem vanishes but when I try to produce a regressor model I get the error message below.
With python 2.7.3 nothing changes still no IPython notebook for me :-(.
Also, it keeps telling me now, that hddm is incompatible with pymc 2.3.6 so it won£t install it...

-Benjamin

m_vat = hddm.HDDMRegressor(data, {"v ~ C(stim, Treatment('congruent'))","a ~ C(stim, Treatment('congruent'))","t ~ C(stim, Treatment('congruent'))"}, group_only_regressors=False, keep_regressor_trace=True, include=['p_outlier'])

Adding these covariates:
['t_Intercept', "t_C(stim, Treatment('congruent'))[T.incongruent]", "t_C(stim, Treatment('congruent'))[T.sound]", "t_C(stim, Treatment('congruent'))[T.word]"]
Adding these covariates:
['a_Intercept', "a_C(stim, Treatment('congruent'))[T.incongruent]", "a_C(stim, Treatment('congruent'))[T.sound]", "a_C(stim, Treatment('congruent'))[T.word]"]
Adding these covariates:
['v_Intercept', "v_C(stim, Treatment('congruent'))[T.incongruent]", "v_C(stim, Treatment('congruent'))[T.sound]", "v_C(stim, Treatment('congruent'))[T.word]"]

---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
<ipython-input-17-d3878d2cd394> in <module>()
----> 1 m_vat = hddm.HDDMRegressor(data, {"v ~ C(stim, Treatment('congruent'))","a ~ C(stim, Treatment('congruent'))","t ~ C(stim, Treatment('congruent'))"}, group_only_regressors=False, keep_regressor_trace=True, include=['p_outlier'])

C:\Anaconda2\lib\site-packages\hddm\models\hddm_regression.pyc in __init__(self, data, models, group_only_regressors, keep_regressor_trace, **kwargs)
    201         self.wfpt_reg_class = deepcopy(wfpt_reg_like)
    202 
--> 203         super(HDDMRegressor, self).__init__(data, **kwargs)
    204 
    205         # Sanity checks

C:\Anaconda2\lib\site-packages\hddm\models\hddm_info.pyc in __init__(self, *args, **kwargs)
    111         self.is_informative = kwargs.pop('informative', True)
    112 
--> 113         super(HDDM, self).__init__(*args, **kwargs)
    114 
    115     def _create_stochastic_knodes(self, include):

C:\Anaconda2\lib\site-packages\hddm\models\base.pyc in __init__(self, data, bias, include, wiener_params, p_outlier, **kwargs)
    687         self.wfpt_class = hddm.likelihoods.generate_wfpt_stochastic_class(wp, cdf_range=self.cdf_range)
    688 
--> 689         super(HDDMBase, self).__init__(data, **kwargs)
    690 
    691     def __getstate__(self):

C:\Anaconda2\lib\site-packages\hddm\models\base.pyc in __init__(self, data, **kwargs)
     38         self.std_depends = kwargs.pop('std_depends', False)
     39 
---> 40         super(AccumulatorModel, self).__init__(data, **kwargs)
     41 
     42 

C:\Anaconda2\lib\site-packages\kabuki\hierarchical.pyc in __init__(self, data, is_group_model, depends_on, trace_subjs, plot_subjs, plot_var, group_only_nodes)
    346         self.db = None
    347 
--> 348         self._setup_model()
    349 
    350     def _setup_model(self):

C:\Anaconda2\lib\site-packages\kabuki\hierarchical.pyc in _setup_model(self)
    357 
    358         # constructs pymc nodes etc and connects them appropriately
--> 359         self.create_model()
    360 
    361     def __getstate__(self):

C:\Anaconda2\lib\site-packages\kabuki\hierarchical.pyc in create_model(self, max_retries)
    431         for tries in range(max_retries):
    432             try:
--> 433                 _create()
    434             except (pm.ZeroProbability, ValueError):
    435                 continue

C:\Anaconda2\lib\site-packages\kabuki\hierarchical.pyc in _create()
    427         def _create():
    428             for knode in self.knodes:
--> 429                 knode.create()
    430 
    431         for tries in range(max_retries):

C:\Anaconda2\lib\site-packages\kabuki\hierarchical.pyc in create(self)
    166                 kwargs['doc'] = node_name
    167 
--> 168             node = self.create_node(node_name, kwargs, grouped_data)
    169 
    170             if node is not None:

C:\Anaconda2\lib\site-packages\kabuki\hierarchical.pyc in create_node(self, node_name, kwargs, data)
    174     def create_node(self, node_name, kwargs, data):
    175         #actually create the node
--> 176         return self.pymc_node(name=node_name, **kwargs)
    177 
    178     def create_tag_and_subj_idx(self, cols, uniq_elem):

C:\Anaconda2\lib\site-packages\pymc\distributions.pyc in __init__(self, *args, **kwds)
    318                     logp_partial_gradients=logp_partial_gradients,
    319                     dtype=dtype,
--> 320                     **arg_dict_out)
    321 
    322     new_class.__name__ = name

C:\Anaconda2\lib\site-packages\pymc\PyMCObjects.pyc in __init__(self, logp, doc, name, parents, random, trace, value, dtype, rseed, observed, cache_depth, plot, verbose, isdata, check_logp, logp_partial_gradients)
    762                           dtype=dtype,
    763                           plot=plot,
--> 764                           verbose=verbose)
    765 
    766         # self._logp.force_compute()

C:\Anaconda2\lib\site-packages\pymc\Node.pyc in __init__(self, doc, name, parents, cache_depth, trace, dtype, plot, verbose)
    212         self.extended_children = set()
    213 
--> 214         Node.__init__(self, doc, name, parents, cache_depth, verbose=verbose)
    215 
    216         if self.dtype is None:

C:\Anaconda2\lib\site-packages\pymc\Node.pyc in __init__(self, doc, name, parents, cache_depth, verbose)
    127 
    128         # Initialize
--> 129         self.parents = parents
    130 
    131     def _get_parents(self):

C:\Anaconda2\lib\site-packages\pymc\Node.pyc in _set_parents(self, new_parents)
    145 
    146         # Get new lazy function
--> 147         self.gen_lazy_function()
    148 
    149     parents = property(

C:\Anaconda2\lib\site-packages\pymc\PyMCObjects.pyc in gen_lazy_function(self)
    811                                       [self]),
    812                                   cache_depth=self._cache_depth)
--> 813         self._logp.force_compute()
    814 
    815         self._logp_partial_gradients = {}

LazyFunction.pyx in pymc.LazyFunction.LazyFunction.force_compute (pymc\LazyFunction.c:2409)()

C:\Anaconda2\lib\site-packages\pymc\distributions.pyc in wrapper(**kwds)
   2979         # Handle Pandas DataFrames
   2980         value = getattr(value, 'values', value)
-> 2981         return f(value, **kwds)
   2982 
   2983     if arguments is None:

C:\Anaconda2\lib\site-packages\hddm\models\hddm_regression.pyc in wiener_multi_like(value, v, sv, a, z, sz, t, st, reg_outcomes, p_outlier)
     27         params = {'v': v, 'sv': sv, 'a': a, 'z': z, 'sz': sz, 't': t, 'st': st}
     28         for reg_outcome in reg_outcomes:
---> 29             params[reg_outcome] = params[reg_outcome].ix[value['rt'].index].values
     30         return hddm.wfpt.wiener_like_multi(value['rt'].values,
     31                                            params['v'], params['sv'], params['a'], params['z'],

IndexError: only integers, slices (`:`), ellipsis (`...`), numpy.newaxis (`None`) and integer or boolean arrays are valid indices

[Thomas Wiecki]

OK, this is a known (and fixed) bug. Just have to release a new version of HDDM which I'll try to do soon.

You can try: pip install git+https://github.com/hddm-devs/hddm and if you're lucky it'll compile.

[stein...@gmail.com]

You are talking about the error from the hddmRegressor I suppose. So should I get rid of the anaconda installation try via pip? Ist that what you are saying? Or what did you mean to suggest?

[Long Ding]

Hi, Benjamin,

It sounds like it could be a version compatibility issue. To run my Windows installation of HDDM, I installed Spyder with python 2.7.3 within the py27_3 environment. Then I activate py27_3 and start Spyder there. 

Long

[Long Ding]

Hi, hddm experts, 

I have two questions from fitting simulated data using HDDMStimCoding and hope you can help me out. 

Here is the construction of the model. Data are attached. 

model = hddm.HDDMStimCoding(data,  stim_col='stim', depends_on={'v':'coh'}, include='z', split_param='v')

Question 1: 

When I ran model.find_starting_values(), I received the following message:

-358.915469175

Traceback (most recent call last):

  File "<ipython-input-42-7d23b9934f61>", line 1, in <module>

    model.find_starting_values()

  File "C:\Anaconda2\envs\py27_3\lib\site-packages\kabuki\hierarchical.py", line 955, in find_starting_values

    self.map(*args, **kwargs)

  File "C:\Anaconda2\envs\py27_3\lib\site-packages\kabuki\hierarchical.py", line 505, in map

    self.draw_from_prior()

  File "C:\Anaconda2\envs\py27_3\lib\site-packages\kabuki\hierarchical.py", line 466, in draw_from_prior

    self.set_values(values)

  File "C:\Anaconda2\envs\py27_3\lib\site-packages\kabuki\hierarchical.py", line 944, in set_values

    self.nodes_db.ix[name]['node'].set_value(value)

  File "C:\Anaconda2\envs\py27_3\lib\site-packages\pymc\PyMCObjects.py", line 478, in set_value

    '\'s value cannot be set.')

AttributeError: Deterministic z's value cannot be set.

Could you please advise what could be causing this? 

Question 2:

Does plot_posterior_predictive() work for models with StimCoding? It seems that, for a given condition, observed RTs are always positive only or negative only, while the predicted values span both positive and negative values. Granted that I was plotting results from not-converged sampling. Just want to understand if the discrepancies in the sign of observed and predicted values are a result or cause of non-convergence. 

Thanks!

Long

[Thomas Wiecki]

Can someone with windows try:

conda install -c pymc hddm

Or, if they already installed hddm:

conda update -c pymc hddm

And let me know if that works. Should install HDDM 0.6.0.

--

[Long Ding]

Hi, Tomas,

I tried the 64-bit Anaconda with Python 3.5

I can import hddm fine and read the csv file (same as the one that I posted earlier in the thread). But I get the following error messages when I try:  

model = hddm.HDDMStimCoding(data,  stim_col='stim', depends_on={'v':['coh']}, split_param='v', include=['z'])

Long

Traceback (most recent call last):

  File "<ipython-input-14-fca813215a2a>", line 1, in <module>

    model = hddm.HDDMStimCoding(data,  stim_col='stim', depends_on={'v':['coh']}, split_param='v', include=['z'])

  File "C:\Users\Long\Anaconda3\lib\site-packages\hddm\models\hddm_stimcoding.py", line 50, in __init__

    super(HDDMStimCoding, self).__init__(*args, **kwargs)

  File "C:\Users\Long\Anaconda3\lib\site-packages\hddm\models\hddm_info.py", line 113, in __init__

    super(HDDM, self).__init__(*args, **kwargs)

  File "C:\Users\Long\Anaconda3\lib\site-packages\hddm\models\base.py", line 689, in __init__

    super(HDDMBase, self).__init__(data, **kwargs)

  File "C:\Users\Long\Anaconda3\lib\site-packages\hddm\models\base.py", line 40, in __init__

    super(AccumulatorModel, self).__init__(data, **kwargs)

  File "C:\Users\Long\Anaconda3\lib\site-packages\kabuki\hierarchical.py", line 348, in __init__

    self._setup_model()

  File "C:\Users\Long\Anaconda3\lib\site-packages\kabuki\hierarchical.py", line 359, in _setup_model

    self.create_model()

  File "C:\Users\Long\Anaconda3\lib\site-packages\kabuki\hierarchical.py", line 433, in create_model

    _create()

  File "C:\Users\Long\Anaconda3\lib\site-packages\kabuki\hierarchical.py", line 429, in _create

    knode.create()

  File "C:\Users\Long\Anaconda3\lib\site-packages\kabuki\hierarchical.py", line 168, in create

    node = self.create_node(node_name, kwargs, grouped_data)

  File "C:\Users\Long\Anaconda3\lib\site-packages\hddm\models\hddm_stimcoding.py", line 108, in create_node

    return self.pymc_node(name, **kwargs)

  File "C:\Users\Long\Anaconda3\lib\site-packages\pymc\distributions.py", line 320, in __init__

    **arg_dict_out)

  File "C:\Users\Long\Anaconda3\lib\site-packages\pymc\PyMCObjects.py", line 764, in __init__

    verbose=verbose)

  File "C:\Users\Long\Anaconda3\lib\site-packages\pymc\Node.py", line 214, in __init__

    Node.__init__(self, doc, name, parents, cache_depth, verbose=verbose)

  File "C:\Users\Long\Anaconda3\lib\site-packages\pymc\Node.py", line 129, in __init__

    self.parents = parents

  File "C:\Users\Long\Anaconda3\lib\site-packages\pymc\Node.py", line 147, in _set_parents

    self.gen_lazy_function()

  File "C:\Users\Long\Anaconda3\lib\site-packages\pymc\PyMCObjects.py", line 813, in gen_lazy_function

    self._logp.force_compute()

  File "LazyFunction.pyx", line 257, in pymc.LazyFunction.LazyFunction.force_compute (pymc\LazyFunction.c:2409)

  File "C:\Users\Long\Anaconda3\lib\site-packages\pymc\distributions.py", line 2981, in wrapper

    return f(value, **kwds)

  File "C:\Users\Long\Anaconda3\lib\site-packages\hddm\likelihoods.py", line 52, in wfpt_like

    return hddm.wfpt.wiener_like(x['rt'].values, v, sv, a, z, sz, t, st, p_outlier=p_outlier, **wp)

[Sam Mathias]

This is the pymc problem again. You'll need to upgrade to pymc 2.5.6 or downgrade to pymc 2.5.4

Samuel Robert Mathias

Postdoctoral Associate
Department of Psychiatry, Yale School of Medicine
Suite 3014, 2 Church Street South
New Haven, CT 06519
http://www.srmathias.com

--

[Thomas Wiecki]

Yes, hopefully.

Can you try

conda update pymc

And report back if that works?

[Long Ding]

There is no pymc 2.5.6 or 2.5.4. Similar error message for both. 

[Anaconda3] C:\Users\Long>conda install pymc=2.5.4

Fetching package metadata: ....

Error: No packages found in current win-64 channels matching: pymc 2.5.4*

You can search for this package on anaconda.org with

    anaconda search -t conda pymc 2.5.4*

In case 2.5.4 is a typo, I tried to install pymc 2.3.4

[Anaconda3] C:\Users\Long>conda install pymc=2.3.4

Fetching package metadata: ....

Solving package specifications: .............

Error: Unsatisfiable package specifications.

Generating hint:

[      COMPLETE      ]|##################################################| 100%

Hint: the following packages conflict with each other:

  - pymc 2.3.4*

  - python 3.5*

Use 'conda info pymc' etc. to see the dependencies for each package.

Note that the following features are enabled:

  - vc14 

Long

[Thomas Wiecki]

On Mon, Dec 7, 2015 at 10:22 PM, Long Ding <longde...@gmail.com> wrote:

conda install pymc=2.3.4

I can't see where you tried

conda install pymc=2.5.6

[Long Ding]

[Anaconda3] C:\Users\Long>conda install pymc=2.5.6

Fetching package metadata: ....

Error: No packages found in current win-64 channels matching: pymc 2.5.6*

You can search for this package on anaconda.org with

    anaconda search -t conda pymc 2.5.6*

[Thomas Wiecki]

Sorry, I meant pymc=2.3.6

--

[Long Ding]

Same message as for pymc=2.3.4 (just replace 4 with 6 in the message). 

[Thomas Wiecki]

OK, it seems like pymc 2.3.6 packages for windows python 3.5 have not been built:

https://github.com/conda/conda/issues/1895

You can try the Python 2.7 version by creating an environment.yml file containing:

name: hddm
dependencies:
- python=2
- pymc

Then, do:

conda env create -f environment.yml

Then:

source activate hddm

Then try to reinstall hddm:

conda install -c pymc hddm

That should pull in HDDM 6.0 for Python 2.7.

[Thomas Wiecki]

OK, pymc=2.3.6 should be added for python 3.5 now. can you try and report back?

[tele...@gmail.com]

Dear Thomas,

What combination of python2/pymc/hddm versions would you currently recommand to have hddm working on a mac and on a linux 64 bit workstation ?

Best,

Philippe

[Thomas Wiecki]

On linux 64 bit I recommend:

python 3, pymc 2.3.6 and hddm 0.6.0. You should get those when you install anaconda3 and do:

conda install -c pymc hddm

I will build OSX packages this weekend so stay tuned.

Thomas