feat: endpoint to rename an attribute name in a table

[Namyalg]

Pull Request Template

Description

This PR adds an endpoint to rename/modify the attribute name of a table

Fixes #177

Type of change

• [ ] Bug fix (non-breaking change which fixes an issue)
• [X] New feature (non-breaking change which adds functionality)

How Has This Been Tested?

The functionality of the endpoint has been tested locally

Working

The endpoint is a POST request that takes in the payload as :

 {
      "title" : "name-of-simulation",
       "tables" : { 
              "table_1 : {
                     "old_name_1" : "new_name_1",
                     "old_name_2" : "new_name_2"
               },

              "table_2 : {
                     "old_name_1" : "new_name_1",
                     "old_name_2" : "new_name_2"
               }

}

Using cURL, it can be performed as :

1. Consider the table turnover_parameter, title of simulation as run4
2. Tableturnover_parameter has the attributes (can be obtained using the endpoint /upload/db/tables ) : ["id","foliage","stems","branch","stem_snag","branch_snag","coarse_root","fine_root","branch_snag_split","coarse_ag_split","fine_ag_split"]

To rename the attribute id to ids, the cURL command is

curl --header "Content-Type: application/json"   --request POST   --data '{"title":"run4","tables" :{"turnover_parameter": {"id" : "ids"}}}'   http://localhost:8080/gcbm/attribute/rename

Erroneous cases handled

1. Case 1 : When a table passed in the payload does not exist Assume the name of the table passed is turn_parameter instead of turnover_parameter, the request will result in

 curl --header "Content-Type: application/json"   --request POST   --data '{"title":"run4","tables" :{"vol_to_bio_factor" : {"a" : "ad"} , "turn_parameter": {"stem" : "stems"}}}'   http://localhost:8080/gcbm/attribute/rename

2. Case 2 : When an attribute name does not exist in a given table

Attribute stems does not exist in turnover_parameter

curl --header "Content-Type: application/json"   --request POST   --data '{"title":"run4","tables" :{"turnover_parameter": {"stems" : "stem"}}}'   http://localhost:8080/gcbm/attribute/rename
{"message":"Error in table turnover_parameter no such column: \"stems\"","status":0}

Questions

In the current design, encountering an error of the 2 types will immediately return from the endpoint with the {status : 0, message : error-that-occured} . This implies, that the changes or queries that were present earlier will be executed.

To provide a robust mechanism, a few other ideas are :

1. Scan through all of the requests, only if there are no possible errors, execute all of the renaming operations
2. Scan through each table, and if a particular table has an error, do not perform any rename operation on that table, while execute the operations on other tables
3. Scan through each table, and if a particular table has an error return an error message immediately, while performing the operations that occur before

What would the best option amongst these ?

@HarshCasper @aornugent @SanjaySinghRajpoot @YashKandalkar I would like to know your opinions

[YashKandalkar]

Another option is to have SQL transactions for renaming attributes. The python sqlite3 library doesn't seem to support transactions by default, but I found this: https://stackoverflow.com/a/23634805/10047268 We can have transactions by this but I think we'll need to manually execute "commit" after each statement in other endpoints if we make the isolation_level to None

[Namyalg]

Thank you @YashKandalkar, will take a look

[HarshCasper]

@Namyalg — Any updates on this? It would be good to resolve the merge conflicts as well.

[Namyalg]

This is not yet resolved

[aornugent]

@Namyalg - do you have time to finish this PR? If not, let's file an issue and try to attract other contributions to finish the featureset.

[Namyalg]

I will work on resolving the issue.