leoek / material-safety-search Goto Github PK

FSC: http://www.wingovernmentcontracts.com/federal-supply-classes.htm
FSG: http://www.wingovernmentcontracts.com/federal-supply-groups-fsg.htm

By adding a mock dataset, developers can try this search without obtaining the original dataset.
The original dataset cannot be uploaded because of licensing issues.

Research and implement processing chains.
E.g. productIdentification field should be lower cased and non alpha-numeric symbols should be removed.

Filter the query for values which are part of the enhanced document and indexed.

@torond link your commit with these indexed fields here.

Add endpoint for query suggestions to server

• Advanced search backend (for each text field suggest fitting names to autocomplete)
• General search word completion

My Suggestion: https://www.gnu.org/licenses/gpl-3.0.html or https://www.gnu.org/licenses/gpl-2.0.html

• Implement try catch block to check if datasheet is present. If it is not, skip the import but leave the server running still.

• Add notion to readme.

• /search endpoint
• /advancedSearch

Some To Dos:

• Change name from 'Material Safety Search' to 'MSDS Search' as that is what we actually do
• Implement Date Selector (both beginDate and endDate must be supplied when wanting to filter)
• Pagination
• Make Title clickable and link to an empty General Search field (Just as one can click on the 'Google' logo to get back to google)
• Logging
• Button for deleting FSG (if possible)
• Put disclaimer somewhere
• Top Terms in General Search

edit Leo:

• sorted by prio
• button for deleting fsg is possible. current workaround is that the search button clears it

default endpoint /search

we have only the parameter s to start with which is a string which will be parsed by the backend. (See #7)

text in the default search bar should be sent to the search endpoint and s parameter exactly as entered.

• /search endpoint is post now and has new parameters

In order to understand needs of the search machines user we should find out what information the user already has and how we can use this information.
Maybe there are differences in different countries.

At the moment only the productIdentification field is being searched for exact matches.
Implement searching on all crucial fields and score documents correctly.

The first version of this auto completion should be based on the most popular searches and the string the string the user already typed.

Currently there is an endpoint /top available which returns the most common search terms. However it does not evaluate any characters which were already typed by the user.

Suggestions for the general search field should be provided with the /suggest endpoint as they are for other fields. Check the openapi spec/swagger for more information.

Add a button or something similiar to switch between the default search bar and the advanced search
Add multiple search fields with autocompletion for:

• company
• cas number
• niin number
• fsc number
• tbc

see #5 for the server part of this ticket.

Should contain:

• Document Response (Page of Documents)
• Info about log endpoint
  To be continued...

Possible ways:
Does every MSDS have the US Disclaimer?
Modify python script to extract countries of origin of MSDS.
NIIN Number contains information about origin.

Integrate SolrJ into current application to unlock advanced Solr features.

Windows, Mac and Linux use different newline symbols.
See class DataSheetImporter for problematic parts.