Fix query API drift and add the continuous MobilityDB query versions

[estebanzimanyi]

Updates the benchmark query scripts to the current MobilityDB API and adds the continuous (MobilityDB) versions of the pattern queries. All four scripts run without error on the Delhi and Paris datasets.

API and script fixes

• atValues(tbool, value) -> atValue for a scalar second argument
• spanUnion(tstzspanset) -> spansetUnion, the aggregate over spansets (current MobilityDB restricts spanUnion to spans)
• close an unterminated block comment (paris_queries.sql)
• disambiguate a duplicate output-column alias (delhi_districts.sql)
• p.BufferGeom -> p.Geom
• comment pasted psql result rows so they parse as comments
• align a DROP/CREATE table name and remove an orphaned empty CREATE TABLE stub
• correct a stale query description (ten minutes)

Continuous query versions

These follow the paper's continuous idiom: whenTrue(<tfloat> #> x) for the exact temporal episode extent, a temporal conjunction with &, and cells via tintSeq(tint(CellId, lower(AtTime))). Each matches its discrete counterpart on the data.

• TGQ5_12 -- episodes where Pm25 > 150 under an always-temperature predicate (8 rows / 4 trips; the continuous approach yields more than the discrete 2 because episodes span cell boundaries)
• TQ5_13 -- Pm25 > 300 and cloudy and humidity > 80 via a temporal conjunction (71 rows / 66 trips)
• Paris part-of-day query over the temporal tour segments (matches the discrete result)