Qiime2 plugins not found though they are installed

Hi I booted an Ubunutu 16.06 instance on Amazon and installed miniconda3 and then followed the instructions available at https://docs.qiime2.org/2.0.6/install/#install-plugins. So, for example, I did this:

conda install matplotlib==1.5.1
conda install -c qiime2 q2-types q2-feature-table

Now if I execute the qiime info command - I get the following that tells me that I don’t have any plugins installed (but I do)

(qiime2) [email protected]:~$ qiime info
System versions
Python version: 3.5.2
QIIME version: 2.0.6
q2cli version: 0.0.6

Installed plugins
No plugins are currently installed.
You can browse the official QIIME 2 plugins at https://qiime2.org

Application config directory
/home/ubuntu/.config/q2cli

So i I try to reinstall the plugins I get the following:

conda install -c qiime2 q2-types q2-feature-table
Fetching package metadata .........
Solving package specifications: ..........

# All requested packages already installed.
# packages in environment at /home/ubuntu/miniconda3/envs/qiime2:
#
q2-feature-table          0.0.6                    py35_0    qiime2
q2-types                  0.0.6                    py35_0    qiime2

Okay let’s repeat qiime info but as before it does not acknowledge the presence of the plugins

(qiime2) [email protected]:~$ qiime info
    System versions
    Python version: 3.5.2
    QIIME version: 2.0.6
    q2cli version: 0.0.6

    Installed plugins
    No plugins are currently installed.
    You can browse the official QIIME 2 plugins at https://qiime2.org

    Application config directory
    /home/ubuntu/.config/q2cli

Also note that there is a problem the first time I do “qiime info” relative to libsm6 which is easily solved by apt-get install libsm - After that I can do the qiime info with no problem. Please advise Thanks

This sounds similar to this issue. I do not recall any errors related to libsm6 — but I had the same issue with qiime not detecting installed plugins on a linux system.

You could attempt this:
echo "backend: Agg" > ~/.config/matplotlib/matplotlibrc

and then see if qiime info lists your plugins as detected.

Hope that helps!

Hi @steviep , just to clarify, you are running the above conda install commands for the plugins in an active conda environment? As in:

[email protected]:~$ source activate qiime2
(qiime2) [email protected]:~$ conda install matplotlib==1.5.1
(qiime2) [email protected]74:~$ conda install -c qiime2 q2-types q2-feature-table

Also, you can run conda list when in a conda environment to view all of your installed packages.

After stewing on this a bit more, the libsm6 warning is related to matplotlib and its default configuration. If you are running this on a headless server (which I assume you are), then you have to configure matplotlib to use a headless rendering backend, just as @Nicholas_Bokulich mentioned. The easiest is to add a config to you local user account that specifies the Agg backend for matplotlib. Just to clarify, the libsm6 is coming from matplotlib and its default configuration, you shouldn’t need that package installed once you configure the Agg backend. Hope that helps!

A post was split to a new topic: qiime tools view on headless server

A post was split to a new topic: Unknown Locale (missing env vars)

Hi, I have the same issue and even after trying - echo “backend: Agg” > ~/.config/matplotlib/matplotlibrc, its still can’t detect the plugins.

You may need to refresh your cache. Try:

qiime dev refresh-cache

and then check for plugins. If that doesn’t fix it, please post back with more details on the exact commands you used to install and any specific error messages, if these deviate from the initial post. Also, to confirm that the plugins are installed correctly, use:

conda list

4 Likes

Thanks , It worked. Now I can see all the installed plugins.

1 Like

Thank you, I have the same problem with the ‘Error: No such command “demux”’ though i have installed it. After the command qiime dev refresh-cache , everything is fine.

2 Likes

Just a quick note that this issue has been resolved, the fix will be available in the next release (February).

1 Like