Clean up of layers in geosafe

[NyakudyaA]

Problem

After playing around with the geosafe I noticed a couple of odd things. They might not be bugs but can be improved.

File duplication

• After running a successful analysis there are results stored in /home/geosafe. The results include summary_table, summary_map_pdf, and shapefile for the results layer.
• The summary_table and summary_pdf is also stored at /usr/src/app/geonode/uploaded/analysis/report The site serves the results from this location.
• After success on the analysis, the shapefile is also copied to /usr/src/app/geonode/qgis_layer where the qgis project file is registered.

NB: Why can we not have a common place where the files are stored to avoid duplication?

Database

• When an analysis is being run it is registered in the Postgres database table geosafe_analysis. The task_state initially starts as pending and should change to success when the analysis finished.
• When the analysis fails a record exists in the database with the status pending regardless of the failed status of the analysis.

select * from geosafe_analysis;
 id |          user_title          |           impact_function_id            | extent_option |               task_id                | task_state | keep | report_map | report_table | aggregation_layer_id | exposur
e_layer_id | hazard_layer_id | impact_layer_id | user_id |                       user_extent                       |           end_time            |          start_time           
----+------------------------------+-----------------------------------------+---------------+--------------------------------------+------------+------+------------+--------------+----------------------+--------
-----------+-----------------+-----------------+---------+---------------------------------------------------------+-------------------------------+-------------------------------
  1 | Nepal earthquake on building | ClassifiedPolygonHazardBuildingFunction |             2 | 2037cd01-c780-4d0f-a053-67d37c42f6f9 | PENDING    | t    |            |              |                      |        
         3 |               1 |                 |    1000 | 85.2333068848,27.7030597274,85.3981018066,27.7769746959 | 2017-09-15 09:43:27.142444+00 | 2017-09-15 09:43:27.142434+00

NB: We should either change the database flag to reflect analysis failed in the database

Solution

Discussion and change when it is free time

[boney-bun]

the save button actives because a user wants to keep the analysis regardless what the result is. the user can simply disable the save button and it will be removed by the end of the day.

yes, the task status should change to failure if an analysis doesn't succeed. we have celery task for updating the status and it run asynchronously. my guess is it happened because of async process. the celery hasn't execute yet by the time you check it. i'm wondering does the task status change in the next day?