The quick and dirty reproducible local setup III

Georgi Iliev · January 22, 2025

Encounter

3. Photo annotation application

The photo annotation application completing the POC solution architecture requires some lightweight configuration to enable the integration with the services we’ve stood up previously.

Check out the code. Assuming you have followed the instructions in the earlier posts, there is no need to modify any parameters already configured in config.py. The missing values are described below.

Pixelfed integration

Find the value for CLIENT_ID and CLIENT_SECRET in the developer settings of your local Pixelfed instance at http://pixelfed.pastabytes.test:8080/settings/developers – this is the client we created when deploying Pixelfed.

CLIENT_ID = ""
CLIENT_SECRET = ""

Virtuoso integration

Set LOCAL_SPARQL_USER to "dba" and LOCAL_SPARQL_PASSWORD to the corresponding secret you already created when running the Virtuoso container.

LOCAL_SPARQL_USER = ""
LOCAL_SPARQL_PASSWORD = ""

Runtime

environment.yaml contains the complete conda environment which should allow you to run the application. Having installed and activated the environment, you can either

1. Run the Streamlit application from the command line

    $ streamlit run app.py
   
        You can now view your Streamlit app in your browser.
   
        Local URL: http://localhost:8501
        Network URL: http://172.23.221.198:8501
   
   

   or

2. Make a Python debug launch configuration in VS Code and run and debug app.py from your IDE. See launch.json below:

    {
    "version": "0.2.0",
        "configurations": [ 
            {
                "name": "Streamlit Debug",
                "type": "debugpy",
                "request": "launch",
                "module": "streamlit",
                "args": [
                    "run",
                    "${file}"
                ]
            }
        ]
    } 
   

That’s it! Provided you have posted at least one picture in your local Pixelfed instance, pointing your browser to the correct alias for your localhost (http://encounter.pixelfed.test:8051) and logging in using the same Pixelfed user should give you your first chance to annotate your photos.

You can inspect the result by running a SPARQL query in the Virtuoso CONDUCTOR SPARQL shell as described previously or in any other SPARQL client such as VSCode SPARQL Notebook (you’ll need to provide your credentials). A successful annotation in the app should produce a triple in response to the query below

    PREFIX encounter-ontology: <https://encounter.pastabytes.com/v0.1.0/ontology/>
    PREFIX skos: <http://www.w3.org/2004/02/skos/core#>

    SELECT ?user, ?bird, ?label WHERE {
        ?s a encounter-ontology:Encounter ;
            encounter-ontology:hasUser ?user ;
            encounter-ontology:hasEvidence ?evidence .
        ?evidence encounter-ontology:depicts ?bird .
        ?bird skos:prefLabel ?label .
        FILTER (lang(?label) = 'en')
        }
    LIMIT 1

user	bird	label
http://pastabytes.test:8080/iliev.georgi	http://www.yso.fi/onto/avio/e2062	hill pigeon@en

Happy birdwatching!

Coming up next

• Product requirements