I'm running into an error that I can't solve despite others having reported the same error.
I am connecting remotely to a Linux machine. I have installed the latest version of anaconda:
$ bash Anaconda2-2.4.0-Linux-x86_64.sh
// A lot of python libraries get installed
installing: _cache-0.0-py27_x0 ...
Python 2.7.10 :: Continuum Analytics, Inc.
creating default environment...
installation finished.
I updated the corresponding paths and it seems like it works:
$ python
Python 2.7.10 |Anaconda 2.4.0 (64-bit)| (default, Oct 19 2015, 18:04:42)
[GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux2
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
Great, so now I want to use conda
, which is pre-installed with Anaconda. It looks like Anaconda gave me 3.18.3:
$ conda --version
conda 3.18.3
Following the test drive instructions, I update conda:
$ conda update conda
Fetching package metadata: An unexpected error has occurred, please consider sending the
following traceback to the conda GitHub issue tracker at:
https://github.com/conda/conda/issues
Include the output of the command 'conda info' in your report.
Traceback (most recent call last):
File "/code/anaconda2-4-0/bin/conda", line 5, in <module>
sys.exit(main())
File "/code/anaconda2-4-0/lib/python2.7/site-packages/conda/cli/main.py", line 195, in main
args_func(args, p)
File "/code/anaconda2-4-0/lib/python2.7/site-packages/conda/cli/main.py", line 202, in args_func
args.func(args, p)
File "/code/anaconda2-4-0/lib/python2.7/site-packages/conda/cli/main_update.py", line 48, in execute
install.install(args, parser, 'update')
File "/code/anaconda2-4-0/lib/python2.7/site-packages/conda/cli/install.py", line 239, in install
offline=args.offline)
File "/code/anaconda2-4-0/lib/python2.7/site-packages/conda/cli/common.py", line 598, in get_index_trap
return get_index(*args, **kwargs)
File "/code/anaconda2-4-0/lib/python2.7/site-packages/conda/api.py", line 42, in get_index
unknown=unknown)
File "/code/anaconda2-4-0/lib/python2.7/site-packages/conda/utils.py", line 119, in __call__
value = self.func(*args, **kw)
File "/code/anaconda2-4-0/lib/python2.7/site-packages/conda/fetch.py", line 237, in fetch_index
session = CondaSession()
File "/code/anaconda2-4-0/lib/python2.7/site-packages/conda/connection.py", line 61, in __init__
super(CondaSession, self).__init__(*args, **kwargs)
File "/usr/lib/python2.7/dist-packages/requests/sessions.py", line 272, in __init__
self.headers = default_headers()
File "/usr/lib/python2.7/dist-packages/requests/utils.py", line 555, in default_headers
'User-Agent': default_user_agent(),
File "/usr/lib/python2.7/dist-packages/requests/utils.py", line 524, in default_user_agent
_implementation = platform.python_implementation()
File "/usr/lib/python2.7/platform.py", line 1521, in python_implementation
return _sys_version()[0]
File "/usr/lib/python2.7/platform.py", line 1486, in _sys_version
repr(sys_version))
ValueError: failed to parse CPython sys.version: '2.7.10 |Anaconda 2.4.0 (64-bit)| (default, Oct 19 2015, 18:04:42) \n[GCC 4.4.7 20120313 (Red Hat 4.4.7-1)]'
Unfortunately, I can't figure out how to avoid this error.
I found a few other StackOverflow posts. This one suggests reinstalling python and pycharm from scratch (but I just installed Anaconda and I don't use pycharm). Another suggests reinstalling canopy, but I'm not using that here. Finally, a third suggests that it's actually a bug, and proposes a fix. Unfortunately, re-naming sys.version
fails to resolve the error. This isn't even my computer so I don't want to get deep into the code and risk messing something up.
I would appreciate some thoughts or advice.
Third solution is correct. This is indeed a bug, since modified sys.version
string breaks a number of platform
module functions that rely on that string having certain format.
However, if you cannot fix it properly, you can attempt a hacky workaround. Module platform
actually has a cache for parsing sys.version
string: so what you have to do is:
platform
module function that unparses sys.version string.Hack itself:
try:
import sys # Just in case
start = sys.version.index('|') # Do we have a modified sys.version?
end = sys.version.index('|', start + 1)
version_bak = sys.version # Backup modified sys.version
sys.version = sys.version.replace(sys.version[start:end+1], '') # Make it legible for platform module
import platform
platform.python_implementation() # Ignore result, we just need cache populated
platform._sys_version_cache[version_bak] = platform._sys_version_cache[sys.version] # Duplicate cache
sys.version = version_bak # Restore modified version string
except ValueError: # Catch .index() method not finding a pipe
pass
You need to put this code somewhere where it will be executed before conda has a chance to fail with exception. Not sure what the best place would be, but you could try it with conda/cli/main.py
, conda/api.py
or conda/connection.py
.