Create announcement about change of URLs

[cmungall]

We should create an announcement on the README about changing URLs

https://s3.amazonaws.com/bbop-sqlite/$ONT.db.gz ==> https://semanticsql.berkeleybop.io/](https://semanticsql.berkeleybop.io/$ONT.db.gz

Note that OAK users will be unaffected if they upgrade >=0.7.2. However, other libraries may directly wrap the old URLs

E.g. https://github.com/vjcitn/ontoProc2

We should also update our own docs e.g
https://github.com/INCATools/semantic-sql/blob/main/notebooks/SemanticSQL-Tutorial.ipynb

[github-actions]

Summary and current state

I checked the repo against this request. Good news: the core fix (serving from a vendor-neutral URL) already landed via #110, and README.md and src/semsql/builder/builder.py already use https://semanticsql.berkeleybop.io/... (with a code comment at builder.py:21-25 explaining the switch and pointing at #110). What's still missing is exactly what this issue asks for:

1. No explicit announcement/migration note in the README. The README just quietly uses the new URL — there's no "if you have s3.amazonaws.com/bbop-sqlite hardcoded, here's what changed and why" callout that downstream users hitting broken links would land on.
2. notebooks/SemanticSQL-Tutorial.ipynb still references the old URL — wget -N https://s3.amazonaws.com/bbop-sqlite/cl.db.gz appears twice (lines 63 and 73). This is the exact file called out in this issue's body.

Related issues (context worth folding into the announcement)

• Provide sqlite via a different URL #110 (closed) — "Provide sqlite via a different URL": this is the PR/issue that actually introduced semanticsql.berkeleybop.io as the vendor-neutral CDN-backed URL. Good to link from the announcement as the "why."
• [bot] Migrate direct raw-S3 bbop-sqlite consumers to the semanticsql CDN (tracking) #115 (open, tracking) — "[bot] Migrate direct raw-S3 bbop-sqlite consumers to the semanticsql CDN." This is essentially the external-consumer half of this same problem, and it already lists vjcitn/ontoProc2 (the exact example cited in this issue) as a hardcoded-URL consumer, alongside biobricks-ai/semsql, several Knowledge-Graph-Hub repos, microbiomedata/ontology-loader, ccb-hms/NHANES-metadata, and monarch-initiative repos. Worth cross-linking Create announcement about change of URLs #121 ↔ [bot] Migrate direct raw-S3 bbop-sqlite consumers to the semanticsql CDN (tracking) #115 since a README announcement is one of the cheapest ways to get those external maintainers' attention. Also notes an important gotcha: the CDN returns 403 to the default Python-urllib/urllib User-Agent (Cloudflare Browser Integrity Check), so anyone hand-swapping the URL also needs to set a non-default User-Agent — worth mentioning explicitly in the announcement so people don't file a second round of "downloads broken" issues.
• Create w3ids for sqlite downloads #43 (open) — "Create w3ids for sqlite downloads": proposes a w3id.org/semsql/ PURL layer so the download URL never needs to change again. If that's still on the roadmap, it might be worth noting in the announcement that a further URL change could happen and a PURL is the long-term fix — or conversely, this could be a good moment to close Create w3ids for sqlite downloads #43 as superseded if semanticsql.berkeleybop.io is now considered stable.
• Use https://s3.amazonaws.com/bbop-sqlite/hp.db.gz as an example S3 download, not https://s3.amazonaws.com/bbop-sqlite/hp.db #79 and Update ontology builds in S3 bucket #76 (both closed) — earlier issues about the old s3.amazonaws.com/bbop-sqlite URLs (correct example file, and stale builds) — purely historical context, no action needed.

Suggested content for the announcement

A short README section (e.g. right after the "Ready-made SQLite3 builds..." line) along the lines of:

Note: as of [date/#110], pre-built databases are served from https://semanticsql.berkeleybop.io/$ONT.db.gz rather than the legacy https://s3.amazonaws.com/bbop-sqlite/$ONT.db.gz. OAK users are unaffected if running OAK ≥ 0.7.2. If you have code that fetches the raw S3 URL directly, please update it — note the CDN requires a non-default HTTP User-Agent (see #115).

Plus fixing the two wget calls in SemanticSQL-Tutorial.ipynb.

I'm not opening a PR per instructions, but if you'd like this actioned automatically:

@dragon-ai-agent please add a "URL change" announcement section to README.md near the existing download-URL mention (after "Ready-made SQLite3 builds..."), explaining the migration from https://s3.amazonaws.com/bbop-sqlite/$ONT.db.gz to https://semanticsql.berkeleybop.io/$ONT.db.gz, noting OAK>=0.7.2 users are unaffected, and noting the CDN requires a non-default User-Agent header (see #115 for the 403/Browser-Integrity-Check gotcha). Also update notebooks/SemanticSQL-Tutorial.ipynb to replace both occurrences of https://s3.amazonaws.com/bbop-sqlite/cl.db.gz with https://semanticsql.berkeleybop.io/cl.db.gz, and cross-link this issue (#121) with #115 and #43 for context.