[hydra] Allow crawling non-data.gouv URLs

[Pierlou]

In local tests, it can be useful to trigger a crawl of URLs that are not of data.gouv resources (for instance a test file from demo). This used to be possible using CLI poetry run analyse-csv --url <url> before URLs/resource_ids where entangled, but now this raises this error (which is expected).
Do we want to change this to allow crawling external URLs just to see how they are handled (csv-detection, parquet, postgres), maybe dropping all results eventually (check, inserted rows...)?