mmisw / cfsn

CF Standard Names Vocabulary Browser

cfsn

A CF Standard Names Vocabulary browser -- prototype

Deployment available at https://mmisw.org/cfsn

See the wiki for documentation.

Trying cfsn locally:

1. Clone the repo git clone https://github.com/mmisw/cfsn.git
2. Make the app/ directory accessible from your web server
3. Open the index file in your browser

Typically you will want or need a local web server for the step 2 above. If you have nodejs installed on your system, you can do the following:

$ npm install http-server -g
$ http-server
Starting up http-server, serving ./ on port: 8080
Hit CTRL-C to stop the server

and then open http://localhost:8080/app/

Noteworthy changes

• 0.1.6: fixing #24

• 0.1.5: fix #23

• 0.1.4: general info in separate lines and other minor adjustments.

• 0.1.3: preliminary mappings included.

• 0.1.2: general info retrieved from endpoint.

• 0.1.1: filtering by category as here but allowing multiple selection and applied as a first filter.

• 0.1.0: global search modes: glob and regex.

• 0.0.5: preliminary inclusion of term links against NVS.

• 0.0.3: options: change page size; showAll button.

• 0.0.2: cache; better hyperlinking in descriptions.

cfsn's Issues

add number of terms _displayed_ to front page

The front page shows the number of terms in the whole vocabulary; it should also show the number displayed given the current search.

Could show this as "Number displayed: ####" (directly under the Number of Terms line), or as "Number of Terms In Vocabulary" vs "Number of Terms Displayed"

filtering on names not working? (depth)

Putting 'depth' into the name filter field (and hitting return) does not exclude some terms that should be excluded.

column search boxes don't support glob/reg exp searching

The column search boxes don't appear to do Glob/Regex searches with Glob or Regex set. (I would assume the setting at the top applies to all searches.)

Add show all button

Could put it next to the number found line in issue #5

capture column search options in url route

Currently (0.0.6), only the global search is captured, eg,
http://mmisw.org/experimental/cfsn/#/search/zenith
would be nice to also capture any search values associated with the individual columns.

Look at related work for good concepts

Just initially capturing interesting pointers provided here ioos/catalog#72 (comment)

on the topic of CF standard_name navigation and parsing, here are some resources (documentation and online systems) I find useful. Some are outdated, but still pretty useful or demonstrate useful ways of skinning the CF SN cat:

• Parsing CF standard names
• Keyword-centred list for CFmetadata Standard Names. Explained in more detail on an email to the CF list.
• CF Standard Names Organized by facets (from EnvThs/CUNY). And an example from this system, listing terms under the "Vector > Wind" facet

faceted search according to categories

Categories as those shown in http://cfconventions.org/Data/cf-standard-names/27/build/cf-standard-name-table.html , but with checkboxes to allow multiple category selection.

What other classification options would be desirable?

glob searching often fails

Using glob search generally now is failing to complete the search. (Only the top search bar appears, nothing below that.) The URL is updated (http://mmisw.org/experimental/cfsn/#/search/!g/) but nothing appears. Just entering that URL appears to duplicate the problem. I thought glob worked at first, now I'm not so sure.

include mappings sections in term view

In the term view include mapping sections (in principle, one per semantic repository and/or SPARQL endpoint) with:

• Mapped terms according to SKOS properties (exactMatch, closeMatch, etc.,), OWL's sameAs property, and others.
• Counts of mappings for the term.

The retrieved URIs will be rendered with hyperlinks as appropriate but further navigation will in general be via direct opening of those links and not through this tool.

Is NVS sparql endopoint CORS enabled?

When making request to http://vocab.nerc.ac.uk/sparql/sparql sometimes a 404 error is reported with the log showing:

No 'Access-Control-Allow-Origin' header is present on the requested resource.

So, check with NERC about actual status of this capability.

make tool search engine friendly

Find some way for this so the tool starts appearing in results when people do google searches for CF names and related info.

Some initial refs:

• http://stackoverflow.com/questions/13499040/how-do-search-engines-deal-with-angularjs-applications
• http://googlewebmastercentral.blogspot.no/2014/05/understanding-web-pages-better.html

Make cell contents break lines even if a long word

CF SN include some loooong terms, which force the cell widths in tables to the maximum term size.

This can be overridden with the CSS style
word-wrap: break-word; (and maybe also overflow-wrap: break-word)
for the cells in question. (see e.g., http://stackoverflow.com/questions/3058866/how-to-force-a-line-break-in-a-loooooong-word-in-a-div for details.)

I think it would be nice to have this as the formatting for our table cells. (Though once I see it I might change my mind....)

regexp search '^age' does not work

Searching on ^age, to find standard variables beginning with 'age', as in http://mmisw.org/experimental/cfsn/#/search/!r/%5Eage, does not find the variable age_of_stratospheric_air (http://mmisw.org/experimental/cfsn/#/age_of_stratospheric_air).

Which is too weird, since the other two age_ variables are found.

poor term Default for non-regexp searching

The term Default is an ambiguous term for indicating non-regular expression searching. It should be something more like 'Plaintext', or 'Literal' or 'Exact'.

separate general info in different lines

(suggested by JG)

Support search by name or by content

Right now search always looks in the definitions, but it would be nice to be able to search by name as well.

missing terms

Tool shows 2489 total terms but the original vocabulary has more.

properly recognize standard names for hyperlinking

Currently this is based on a simple regex (lowercase letters with embedded underscores), which does apply to most real names but skip others, and creates false positives, etc.

run well also if served from https

display metadata including version information

In particular, links to appropriate vocabulary sources, and associated "version_number" ; also the version of the ontology representation.

canonical units links

The idea here is to provide link(s) from this field to the associated units definition(s) in some semantic repository.

In general, we would probably need to parse the string according to the unit grammar to properly extract the relevant components for a comprehensive association of links. However, unless there is already some Javascript library for this purpose, we can simply tokenize the string and provide links only for the atomic components that have corresponding symbols in the UDUNITS vocabularies.

Refs:

• https://github.com/graybealski/cf-conventions-work/blob/master/FAQ.md#how-do-i-specify-units-in-cf
• http://www.unidata.ucar.edu/software/udunits/udunits-2.2.16/doc/udunits/udunits2lib.html#Examples

consider supporting a regexp or glob search

[suggestion received via email]

fix field rendering in firefox

and possibly other browsers (testing so far has only been on Chrome), for example, on firefox 30.0 on a Mac the search and page fields are not displayed correctly: