fix(sql-schema): stop generating a duplicate "id" column

[albertoabellagarcia]

Summary

Found while verifying #94 (anyOf support, merged): generate_sql_schema() was generating a duplicate "id" column in effectively every schema it produces.

id properties almost always declare their own "type" (e.g. "type": "string"), so the main type-mapping branch already appends a column for id. Then the unconditional if key == "id": block at the end of the loop appends id again, this time as TEXT PRIMARY KEY -- with no check for whether it was already added above. Every generated schema.sql with an id property ends up with something like:

CREATE TABLE AirQualityObserved ("address" JSON, ..., "id" TEXT, ..., "id" TEXT PRIMARY KEY, ...);

which is invalid SQL -- PostgreSQL rejects CREATE TABLE statements that declare the same column twice.

Fix

Moved the id special-case to the top of the loop, with a continue: id is now handled exactly once and always typed TEXT PRIMARY KEY, which was already the effective outcome (the second append was overriding the first one's type) -- this just stops the first append from happening at all for id.

Testing

• Verified against AirQualityObserved (the model from generate_sql_schema - support for anyOf #87): "id" now appears exactly once in the generated schema, correctly typed as TEXT PRIMARY KEY.
• Re-ran generate_sql_schema() against all 231 real model.yaml files available in a local FIWARE GitHub org checkout: no exceptions, and exactly one "id" column for every entity that declares an id property. 3 entities in that set (Distribution, Presentation) don't declare an id property at all and are correctly unaffected.

Independent of #87/#94 -- this bug exists for any model with an id property, regardless of whether it also uses anyOf.