unique URL for each query + meta data

[waehlisch]

Is your feature request related to a problem? Please describe.
for documentation and reproducibility reasons, it is important (i) to replay the query that was executed to produce specific results and (ii) to get an overview of the IYP system state when the query was executed, in particular which data sources were used.

such feature would ease referencing results produced by IYP, which is important in research papers.

Describe the solution you'd like
a unique URL that refers to a specific query executed, as well as unique URL that groups multiple queries.

based on keywords in the URL, the URL redirects to the browser to replay the query (e.g., https://iyp.iijlab.net/iyp/browser/...) or to an overview of the data used (https://iyp.iijlab.net/iyp/data/...). the data page should list all data sources used as well as additional meta data (e.g., timestamps) of the data sources.

Describe alternatives you've considered
the current web frontend allows to save queries as favorites. favorites can be grouped into folders. favorites, including folders, can be downloaded as scripts or zip archive. this allows to create artefacts and store them somewhere else -- maybe a hint in the documentation would be good. this approach is, however, less convenient than proposed above. more importantly, the IYP system state when the query was executed is not recorded.

Additional context
triggered by discussions during the CAIDA GMI-AIMS-5 Workshop, and inspired by a feature @amaennel plans for UCSD-NT.

[m-appel]

Hi Matthias, thank you for your feedback!

We will probably divide this into smaller sub-issues, since there are some parts we plan to address soon, but others which will take some time.

Reproducibility

Since the current frontend is Neo4j Browser, we can not easily modify it, i.e., providing a replay functionality with a unique link is not possible at the moment. However, there are plans to implement our own frontend that would enable something like this (long term).

That being said, we do plan to add documentation explicitly for researchers that explains what and how things should be documented to enable reproduction. At the moment this would include specifying the dump that was used and gathering all queries in a Jupyter notebook or the like.

Status page

We got this feedback from multiple people at AIMS, so we plan to create a page that makes it easier to view which data is available in both the currently deployed version, as well as in past dumps.

I also took note of your comments regarding the favorites feature of the Neo4j Browser. We will add it to the documentation, but it is indeed not a good replacement, since the queries are stored in the browser data, i.e., if you switch machines or have a privacy-sensitive configuration that deletes this data on close, the queries are also gone...

[dbasu03]

Hi, can I solve this issue or is it already assigned?

[m-appel]

Since this is a larger piece of work, we are currently considering it as part of our GSoC ideas:

• IYP browser
• IHR status page

Feel free to checkout and participate in the discussions.

[HUDASULTANA]

@m-appel Would it make sense to include a short hash (like truncated SHA256) of the query + system state in the generated URL? That way even if URLs changes later still reproducibility is easier to verify, the hash could also be shown in the results page footer for citation purposes.