Users should be able to navigate through the database by double-clicking on any technosphere exchange. (And perhaps get additional information on the biosphere exchanges by clicking on them as well, e.g. CFs? not sure what exactly at this point).

test log output

Log Output:
```Filters will effect all database queries until unset (.filters = None)
Using the following database filters:
{'type': 'process'}
Using the following database filters:
{'type': 'process'}
Filters will effect all database queries until unset (`.filters = None`)
Using the following database filters:
{'type': 'emission'}
Using the following database filters:
{'type': 'emission'}
Filters will effect all database queries until unset (`.filters = None`)
Using the following database filters:
{'type': 'emission'}
Using the following database filters:
{'type': 'emission'}

Because the import tasks run in a separate thread, they continue running even if the import is "canceled". See also review comment in #41 .

see the issue on bitbucket

• Right now it's possible to open two wizards which leads to various crashes, eg orphaned QThreads that segfault and things like that
• If there is already a database with db_name, it is silently overwritten and the UI doesn't display the progress of writing to sqlite:
  
• If the users selects a directory without ecospold files, there is no useful feedback and the wizard looks like this:
  

There is quite a bit functionality missing for creating new activities:

• defining the reference product
• defining the unit
• defining uncertainties (e.g. pedigree matrix), probably an issue by itself
• ... to be completed

How can process/elementary flow contributions to LCA results be communicated in a graphically compelling way (e.g. interactive Sankey diagrams displaying inventory flows or impact category "flows" along the supply chain)?
This is closely linked to issue #22.
A widget should be developed that enables:
• Display (parts of) the supply chain (as required by the user)
• Enable a graphical supply chain navigation
• Cut-offs to focus on the relevant part of the supply chains (defined by flow/impact relevance or the interest of the user)

Certain databases, like ecoinvent should be read-only, or in other words should be protected so the user cannot accidentally mess them up.
The Activity Browser currently already prevents activities from being deleted when other activities still use them, but, what is not prevented, e.g. is that exchanges can be added to activities by dragging and dropping...

Test bug report from CML Leiden.

Log Output:

Brightway2 data directory: C:\Users\steubingbrp\AppData\Local\pylca\Brightway3
Brightway2 active project: default
Switched to C:\Users\steubingbrp\surfdrive\Leiden\python\Brightway3 as Brightway2 data directory.
Switched to Brightway2 data directory: C:\Users\steubingbrp\surfdrive\Leiden\python\Brightway3
Filters will effect all database queries until unset (`.filters = None`)
Using the following database filters:
{'type': 'process'}
Using the following database filters:
{'type': 'process'}
Filters will effect all database queries until unset (`.filters = None`)
Using the following database filters:
{'type': 'emission'}
Using the following database filters:
{'type': 'emission'}
Filters will effect all database queries until unset (`.filters = None`)
Using the following database filters:
{'type': 'emission'}
Using the following database filters:
{'type': 'emission'}

As promised I start to contribute....or bother depending on the point of view :P

LCA Calculations does not work when only one FU is used (regardless of the number of methods), the prob does no occurs with more than one.
See error below

Entering clear layout
Total: 0
Traceback (most recent call last):
  File "/home/giuseppec/miniconda3/envs/param/lib/python3.6/site-packages/lca_activity_browser/app/ui/tabs/lca_results.py", line 82, in calculate
    process_contribution_plot = LCAProcessContributionPlot(self, self.mlca)
  File "/home/giuseppec/miniconda3/envs/param/lib/python3.6/site-packages/lca_activity_browser/app/ui/graphics.py", line 139, in __init__
    df_tc = pd.DataFrame(tc)
  File "/home/giuseppec/miniconda3/envs/param/lib/python3.6/site-packages/pandas/core/frame.py", line 330, in __init__
    mgr = self._init_dict(data, index, columns, dtype=dtype)
  File "/home/giuseppec/miniconda3/envs/param/lib/python3.6/site-packages/pandas/core/frame.py", line 461, in _init_dict
    return _arrays_to_mgr(arrays, data_names, index, columns, dtype=dtype)
  File "/home/giuseppec/miniconda3/envs/param/lib/python3.6/site-packages/pandas/core/frame.py", line 6130, in _arrays_to_mgr
    index = extract_index(arrays)
  File "/home/giuseppec/miniconda3/envs/param/lib/python3.6/site-packages/pandas/core/frame.py", line 6173, in extract_index
    index = _union_indexes(indexes)
  File "/home/giuseppec/miniconda3/envs/param/lib/python3.6/site-packages/pandas/core/indexes/api.py", line 67, in _union_indexes
    result = Index(sorted(result))
TypeError: '<' not supported between instances of 'tuple' and 'str'

Entering clear layout
Total: 0
Traceback (most recent call last):
  File "/home/giuseppec/miniconda3/envs/param/lib/python3.6/site-packages/lca_activity_browser/app/ui/tabs/lca_results.py", line 82, in calculate
    process_contribution_plot = LCAProcessContributionPlot(self, self.mlca)
  File "/home/giuseppec/miniconda3/envs/param/lib/python3.6/site-packages/lca_activity_browser/app/ui/graphics.py", line 139, in __init__
    df_tc = pd.DataFrame(tc)
  File "/home/giuseppec/miniconda3/envs/param/lib/python3.6/site-packages/pandas/core/frame.py", line 330, in __init__
    mgr = self._init_dict(data, index, columns, dtype=dtype)
  File "/home/giuseppec/miniconda3/envs/param/lib/python3.6/site-packages/pandas/core/frame.py", line 461, in _init_dict
    return _arrays_to_mgr(arrays, data_names, index, columns, dtype=dtype)
  File "/home/giuseppec/miniconda3/envs/param/lib/python3.6/site-packages/pandas/core/frame.py", line 6130, in _arrays_to_mgr
    index = extract_index(arrays)
  File "/home/giuseppec/miniconda3/envs/param/lib/python3.6/site-packages/pandas/core/frame.py", line 6173, in extract_index
    index = _union_indexes(indexes)
  File "/home/giuseppec/miniconda3/envs/param/lib/python3.6/site-packages/pandas/core/indexes/api.py", line 67, in _union_indexes
    result = Index(sorted(result))
TypeError: '<' not supported between instances of 'tuple' and 'str'

Test QInputDialog.getMultilineText

this is multiline text

What about markdown?

google

Log Output:

Brightway2 data directory: /home/adrian/.local/share/Brightway3
Brightway2 active project: default

It would be nice to have some automated tests with a CI-service. First of all just to make sure that the activity-browser actually runs on all platforms to catch things like the issue with QtWebEngineWidgets earlier (see #16 ). Secondly it would also be great to have some aspects of the GUI tested automatically, for example with pytest-qt.

this is a test

The javascript-files need to be packaged with the python files for the sankey visualization to work:

• add the required files to setup.py to make pip install https://github.com/LCA-ActivityBrowser/activity-browser/archive/master.zip work
• add the files to meta.yaml to make it work with conda install ...

See screenshot below:

Layout of activity tab needs a complete overhaul. See also metion of this issue in #7.

Required developments:
A robust data entry mask for the AB including
• Economic and environmental exchanges
• Unit, location, etc.
• Uncertainties
• Possibly parameters
• Copying, saving, and deleting activities from selected databases
etc.

CS only work if the project is reloaded... also, at start the CS shows 1,2,3 in "Products and amounts" as well as the 4 buttons (CS selection, new, rename, delete), but one cannot select from CS selection. It works once the project has been reloaded...

At the moment my linter-plugin is complaining whenever I open one of the py-files. Need to go through the code base make it more pep8 conforming.

HI guys,
when I open a database in a project, while I can open activities in the separate tab this is not possible with biosphere datasets. Is this the expected behaviour.
I think would be convenient to be able also to open such a type e.g. to explore the activities that uses a certain biosphere flow downstream.

This is now a test from the activity browser itself

Users should be able to choose their database directory from within the GUI.
Currently, there is a workaround, by creating a app/settings.py file with two variables:

BW2_DIR = r'YOUR PATH' # the directory that your brightway2 data(bases) are located in
PROJECT_NAME = 'YOUR PROJECT NAME' # the project that you'd like to open when the activity-browser starts

The os selectors in the conda build recipe don't do what I expected them to do:

- pyqt # [osx]
- pyqt >=5.9 # [not osx]

See explanation here: conda/conda-build#1024

Maybe the easiest fix for this is to build pyqt 5.9 for OS X. Otherwise the dev version noarch build on travis won't work, no matter what.

How to display LCA results in informative and graphically appealing ways? Here some ideas:
Different plots
• Inventory level (economic/environmental flows)
• LCA score(s) (Barchart, Piechart, Treemap…)
• Contribution Analyses (by process, elementary flow, product, geography, ISIC sector) (Barchart, Treemap, Piechart, Worldmap for geo)
• Uncertainties: Monte Carlo (possibly also others, e.g. Latin-Hypercube) (needs also to integrated in to calculation setups)
Tables to export the data
• To be developed first
User choices:
• LCIA method, cut-off

This is currently not working and too prominent. Perhaps we could have it report issues to this repository, but in the MenuBar under "Help".

HI Hadrian,
cool that you took this project over, do you have any development plan perhaps?
I think the development of a GUI would tremendously help all the bw2 users.....I guess almost everyone is using bw2 for more advanced things but still, most of the time is often spent in working on the databases, and having GUI would allow to save a tremendous amount of time for doing other things....
I am willing to help...even if I have basically no experience with pyqt...

To make the activity-browser easily accessible to non-python users, a standalone executable would be required. The most promising way to do this is with pyinstaller, especially because it allows to build self-contained python applications for all major platforms.

A standalone executable would also make it necessary for the activity-browser to have the option to import databases from raw data (ie. ecoinvent db from ecospold).

Test new implementation of raising issues directly

Log Output:

Brightway2 data directory: /home/adrian/.local/share/Brightway3
Brightway2 active project: default

Another test bug report from Leiden.

Log Output:

Brightway2 data directory: C:\Users\steubingbrp\AppData\Local\pylca\Brightway3
Brightway2 active project: default
Switched to C:\Users\steubingbrp\surfdrive\Leiden\python\Brightway3 as Brightway2 data directory.
Switched to Brightway2 data directory: C:\Users\steubingbrp\surfdrive\Leiden\python\Brightway3
Filters will effect all database queries until unset (`.filters = None`)
Using the following database filters:
{'type': 'process'}
Using the following database filters:
{'type': 'process'}
Filters will effect all database queries until unset (`.filters = None`)
Using the following database filters:
{'type': 'emission'}
Using the following database filters:
{'type': 'emission'}
Filters will effect all database queries until unset (`.filters = None`)
Using the following database filters:
{'type': 'emission'}
Using the following database filters:
{'type': 'emission'}
Test bug report from CML Leiden.

Log Output:

Brightway2 data directory: C:\Users\steubingbrp\AppData\Local\pylca\Brightway3
Brightway2 active project: default
Switched to C:\Users\steubingbrp\surfdrive\Leiden\python\Brightway3 as Brightway2 data directory.
Switched to Brightway2 data directory: C:\Users\steubingbrp\surfdrive\Leiden\python\Brightway3
Filters will effect all database queries until unset (.filters = None)
Using the following database filters:
{'type': 'process'}
Using the following database filters:
{'type': 'process'}
Filters will effect all database queries until unset (.filters = None)
Using the following database filters:
{'type': 'emission'}
Using the following database filters:
{'type': 'emission'}
Filters will effect all database queries until unset (.filters = None)
Using the following database filters:
{'type': 'emission'}
Using the following database filters:
{'type': 'emission'}

test if log is appended to message

Log Output:
<PyQt5.QtWidgets.QTextEdit object at 0x7f67c671c708>

Time is wrong by 1 hour:

Installation instructions should be added to the README

pep8

Guidelines are to follow this order:

• standard lib
• 3d party module
• local imports

separated by blank line

Split out the relevant files for conda build to their own conda-recipe repo. Otherwise the whole repo is included in the recipe.

Hi guys,
first happy to see you take the development back.

I have just installed AB in an already created environment and some packages got downgraded

The following NEW packages will be INSTALLED:

    activity-browser: 2.2.1-pyhd50db26_1 haasad     

The following packages will be UPDATED:

    pyqt:             5.6.0-py36_4       conda-forge --> 5.9.2-py36h7ca782d_0  haasad
    sip:              4.18-py36_1        conda-forge --> 4.19.6-py36h7b295a5_1 haasad

The following packages will be SUPERSEDED by a higher-priority channel:

    fontconfig:       2.12.1-4           conda-forge --> 2.11.1-6                    
    freetype:         2.7-1              conda-forge --> 2.5.5-2                     
    matplotlib:       2.1.0-py36_0       conda-forge --> 2.0.2-np113py36_0           
    qt:               5.6.2-3            conda-forge --> 4.8.7-3                     

Dunno if this unwanted behaviour or is necessary to make AB works, but in the latter case I would specify in the sectionInstall the activity browser of the README that AB must/should be installed in a separate environment to avoid this.

The history tab is not populated

What is the best graphical representation / interactive visualization to understand complex supply chain networks as described in life cycle inventory databases?
This should be implemented in a widget that is able to:
• Display (parts of) the supply chain (as required by the user)
• Enable graphical supply chain navigation (i.e. by right/left-clicking on processes, and using buttons, etc., the user can comfortably navigate through the supply chain and get a quick understanding of it)

File "/Users/cmutel/miniconda3/envs/isie/lib/python3.6/site-packages/lca_activity_browser/app/ui/panels/left.py", line 5, in <module> from ..graphics import DefaultGraph File "/Users/cmutel/miniconda3/envs/isie/lib/python3.6/site-packages/lca_activity_browser/app/ui/graphics.py", line 9, in <module> from seaborn.linearmodels import corrplot ImportError: cannot import name 'corrplot'

Issue on bitbucket

The vertical layout of the activity tab needs to be improved to use the space more efficiently:

Things to improve:

• Reduce margin space / unused vertical space
• Resize tables based on content
• Display comment only when triggerd by some event (eg. button, or hover tooltip)

From original issue on bitbucket:

• Allow tables to expand/collapse with more details, e.g. database, parameterized exchange, uncertainty type, uncertainty values

• Who exactly is going to be the copyright holder now?
• Use github's template to make sure license type is display in header of repo (see picture)
  

The main tabs can't be closed and therefore shouldn't have the option available:

Ideally new, closable tabs should be opened as subtabs of a non-closeable main tab (subtab example from another app):

Like this the main tabs aren't cluttered by opening many activity tabs:

The search function is very basic and could be much improved.

So far the activity browser is only installable via conda:
conda install -c conda-forge -c cmutel -c haasad activity-browser

Pip install is only possible via github:
pip install https://github.com/haasad/activity-browser/archive/master.zip

install_requires in setup.py needs to be updated to make github pip install work properly.

sometimes there is no description at all and then this widget takes up unnecessary space...

While it works for new install/ new environments, conda doesn't understand that the package has changed because there is no increase in build nor version number.

Because of this conda update activity-browser-dev doesn't update atm.

See discussion in issue here
Straightforward installation with conda is not possible at the moment on linux because of this. Need to verify if it works on Windows without installing pip dependencies

In order to explore a database with the activity browser it would be nice to have an advanced search feature, where one can filter for different attributes:

• name, search string contained in name
  • maybe even fuzzy matching with fuzzywuzzy
• activity type
• location
• reference product
• economic sector (via ISIC code)
• biosphere flows (see #3 )

Below is a screenshot of a tool with some of these capabilities that I built some time ago:

This is just a dirty trick taken from here: http://solutionoptimist.com/2013/12/28/awesome-github-tricks/

Like this the issue thread can be abused to upload/host the images without cluttering the repo

We need to come up with a system to plan the development. The current CHANGES.txt file is not the best way forward in my opinion. I have no experience yet with the github projects and milestones, but it appears to work very well in other projects.

I think it would be useful to translate things mentioned in #1 and also some of the very broad issues like #21, #22 and #23 to projects and milestones or something similar. I think an issue should ideally be small and centered around one problem that can be resolved with a single pull-request (or a more general discussion like here).

Other projects use certain milestones to define when the next release is ready. We are now at the somewhat random version number 2.2.1. I think it makes sense to tie the stable release of the activity-browser package on conda to a version number. Additionally the most up-to-date state of AB is available as activity-browser-dev (with version number 2.2.1.dev).

Happy to hear any inputs/suggestions/opinions

This might be related to @bsteubing changes with setFullscreen, since the behaviour of fullscreen appears to differ between operating systems. Need to investigate further.