Error running models (pytensor?)

[cmkaplan]

Hi, and thank you in advance for helping me with this. I am brand new to HDDM/HSSM (exploring a switch over from R Stan Wiener/Diffusion modeling). I use an M1 Mac and could not get HDDM installed, so I am trying to implement HSSM instead. I successfully installed the HSSM package in a python 3.11 anaconda environment. I am able to import hssm. But when I try to run any model code at all (e.g., simple_ddm_model = hssm.HSSM(data=dataset) from the introduction tutorial), I get a lengthy error culminating in the following:

Traceback (most recent call last):

  Cell In[22], line 1
    simple_ddm_model = hssm.HSSM(data=dataset)

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/hssm/hssm.py:315 in __init__
    self.model_distribution = self._make_model_distribution()

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/hssm/hssm.py:1103 in _make_model_distribution
    return make_distribution(

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/hssm/distribution_utils/dist.py:399 in make_distribution
    lapse_func = pytensor.function(

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/compile/function/__init__.py:315 in function
    fn = pfunc(

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/compile/function/pfunc.py:468 in pfunc
    return orig_function(

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/compile/function/types.py:1756 in orig_function
    fn = m.create(defaults)

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/compile/function/types.py:1646 in create
    start_import_time = pytensor.link.c.cmodule.import_time

AttributeError: module 'pytensor' has no attribute 'link'

This suggests an issue with the pytensor package? My current version is 2.17.3

Any assistance would be greatly appreciated. I am so eager to get this up and running but have kept having issues. Happy to send any more information requested as needed to help solve this.

[cmkaplan]

As an update, I am able to run the tutorial sample model code in a python terminal window and jupyter notebook just fine. The problem seems to be with Spyder. I was able to overcome the prior error and now am able to run the "simple_ddm_model = hssm.HSSM(data=dataset)" code with simulated data without issue. I am able to specify the example model based on the cav_data, but when I try to execute "model.sample()" I get a whole new error (again, just in Spyder):

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/link/c/cmodule.py:2649 in compile_str
    raise CompileError(

CompileError: Compilation failed (return status=1):
/usr/bin/clang++ -dynamiclib -g -O3 -fno-math-errno -Wno-unused-label -Wno-unused-variable -Wno-write-strings -Wno-c++11-narrowing -fno-exceptions -fno-unwind-tables -fno-asynchronous-unwind-tables -DNPY_NO_DEPRECATED_API=NPY_1_7_API_VERSION -fPIC -undefined dynamic_lookup -I/opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/numpy/core/include -I/opt/anaconda3/envs/hssm311/include/python3.11 -I/opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/link/c/c_code -L/opt/anaconda3/envs/hssm311/lib -fvisibility=hidden -o /Users/cmkaplan/.pytensor/compiledir_macOS-14.1.1-arm64-arm-64bit-arm-3.11.8-64/tmpc5mlwklv/me70326816b591dfc9dcda2dd5cf701329350d27441c3f1f174f02fb219972685.so /Users/cmkaplan/.pytensor/compiledir_macOS-14.1.1-arm64-arm-64bit-arm-3.11.8-64/tmpc5mlwklv/mod.cpp
clang: error: unable to make temporary file: No such file or directory



During handling of the above exception, another exception occurred:

Traceback (most recent call last):

  Cell In[7], line 1
    model.sample()

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/hssm/hssm.py:415 in sample
    self._inference_obj = self.model.fit(

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/bambi/models.py:325 in fit
    return self.backend.run(

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/bambi/backend/pymc.py:96 in run
    result = self._run_mcmc(

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/bambi/backend/pymc.py:172 in _run_mcmc
    idata = pm.sample(

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pymc/sampling/mcmc.py:689 in sample
    step = assign_step_methods(model, step, methods=pm.STEP_METHODS, step_kwargs=kwargs)

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pymc/sampling/mcmc.py:239 in assign_step_methods
    return instantiate_steppers(model, steps, selected_steps, step_kwargs)

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pymc/sampling/mcmc.py:140 in instantiate_steppers
    step = step_class(vars=vars, model=model, **args)

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pymc/step_methods/hmc/nuts.py:180 in __init__
    super().__init__(vars, **kwargs)

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pymc/step_methods/hmc/base_hmc.py:109 in __init__
    super().__init__(vars, blocked=blocked, model=self._model, dtype=dtype, **pytensor_kwargs)

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pymc/step_methods/arraystep.py:164 in __init__
    func = model.logp_dlogp_function(vars, dtype=dtype, **pytensor_kwargs)

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pymc/model/core.py:612 in logp_dlogp_function
    return ValueGradFunction(costs, grad_vars, extra_vars_and_values, **kwargs)

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pymc/model/core.py:351 in __init__
    self._pytensor_function = compile_pymc(inputs, outputs, givens=givens, **kwargs)

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pymc/pytensorf.py:1195 in compile_pymc
    pytensor_function = pytensor.function(

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/compile/function/__init__.py:315 in function
    fn = pfunc(

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/compile/function/pfunc.py:468 in pfunc
    return orig_function(

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/compile/function/types.py:1756 in orig_function
    fn = m.create(defaults)

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/compile/function/types.py:1649 in create
    _fn, _i, _o = self.linker.make_thunk(

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/link/basic.py:254 in make_thunk
    return self.make_all(

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/link/vm.py:1252 in make_all
    raise_with_op(fgraph, node)

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/link/utils.py:535 in raise_with_op
    raise exc_value.with_traceback(exc_trace)

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/link/vm.py:1243 in make_all
    node.op.make_thunk(node, storage_map, compute_map, [], impl=impl)

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/link/c/op.py:131 in make_thunk
    return self.make_c_thunk(node, storage_map, compute_map, no_recycling)

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/link/c/op.py:96 in make_c_thunk
    outputs = cl.make_thunk(

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/link/c/basic.py:1200 in make_thunk
    cthunk, module, in_storage, out_storage, error_storage = self.__compile__(

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/link/c/basic.py:1120 in __compile__
    thunk, module = self.cthunk_factory(

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/link/c/basic.py:1644 in cthunk_factory
    module = cache.module_from_key(key=key, lnk=self)

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/link/c/cmodule.py:1240 in module_from_key
    module = lnk.compile_cmodule(location)

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/link/c/basic.py:1543 in compile_cmodule
    module = c_compiler.compile_str(

  File /opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/link/c/cmodule.py:2649 in compile_str
    raise CompileError(

CompileError: Compilation failed (return status=1):
/usr/bin/clang++ -dynamiclib -g -O3 -fno-math-errno -Wno-unused-label -Wno-unused-variable -Wno-write-strings -Wno-c++11-narrowing -fno-exceptions -fno-unwind-tables -fno-asynchronous-unwind-tables -DNPY_NO_DEPRECATED_API=NPY_1_7_API_VERSION -fPIC -undefined dynamic_lookup -I/opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/numpy/core/include -I/opt/anaconda3/envs/hssm311/include/python3.11 -I/opt/anaconda3/envs/hssm311/lib/python3.11/site-packages/pytensor/link/c/c_code -L/opt/anaconda3/envs/hssm311/lib -fvisibility=hidden -o /Users/cmkaplan/.pytensor/compiledir_macOS-14.1.1-arm64-arm-64bit-arm-3.11.8-64/tmpc5mlwklv/me70326816b591dfc9dcda2dd5cf701329350d27441c3f1f174f02fb219972685.so /Users/cmkaplan/.pytensor/compiledir_macOS-14.1.1-arm64-arm-64bit-arm-3.11.8-64/tmpc5mlwklv/mod.cpp
clang: error: unable to make temporary file: No such file or directory

Apply node that caused the error: Composite{...}([[  -1.] [ ... ] [-400.]], ExpandDims{axis=0}.0, ExpandDims{axes=[0, 1]}.0, [[ 1.] [ 2 ... 9.] [20.]], ExpandDims{axis=0}.0)
Toposort index: 36
Inputs types: [TensorType(float32, shape=(20, 1)), TensorType(float32, shape=(1, 3988)), TensorType(float32, shape=(1, 1)), TensorType(float32, shape=(20, 1)), TensorType(float32, shape=(1, None))]

HINT: Use a linker other than the C linker to print the inputs' shapes and strides.
HINT: Re-running with most PyTensor optimizations disabled could provide a back-trace showing when this node was created. This can be done by setting the PyTensor flag 'optimizer=fast_compile'. If that does not work, PyTensor optimizations can be disabled with 'optimizer=None'.
HINT: Use the PyTensor flag `exception_verbosity=high` for a debug print-out and storage map footprint of this Apply node.

[digicosmos86]

Hi @cmkaplan,

This seems to happen when pytensor tries to compile the computation graph for the likelihood computation. The issue is that clang cannot make temporary files. Since this issue only happens in Spyder, I think it's a permission issue specific to Spyder. Can you check if Spyder has access to the .pytensor directory under your home directory? The terminal that's running in Spyder should also have access