feat(meos): add trgeometry vs geometry spatial predicate NES operators (W148)#168
Open
estebanzimanyi wants to merge 97 commits into
Open
Conversation
…aho-mqtt-cpp The Nix environment provides paho-mqtt-cpp as a shared-only library and does not include libmeos. Four changes make the build pass: - flake.nix: add stock paho-mqtt-c + paho-mqtt-cpp to baseThirdPartyDeps; pass -DNES_ENABLE_MEOS=OFF in both defaultPackage and devShell cmakeFlags. - nes-plugins/CMakeLists.txt: introduce NES_ENABLE_MQTT and NES_ENABLE_MEOS options that gate the respective plugin subdirectories via activate_optional_plugin(). - nes-plugins/Sources/MQTTSource, nes-plugins/Sinks/MQTTSink: select PahoMqttCpp::paho-mqttpp3-static when available, fall back to the shared PahoMqttCpp::paho-mqttpp3 target (Nix only ships the shared variant). - nes-physical-operators: gate all MEOS-specific add_plugin calls and the nes-meos link behind if(NES_ENABLE_MEOS) in three CMakeLists files (top-level, Functions/Meos, Aggregation/Function/Meos). - CMakeLists.txt (root): declare option(NES_ENABLE_MEOS) and add_compile_definitions(NES_ENABLE_MEOS) before the nes-* subdirectory loop so the preprocessor symbol is visible to all sibling components. - nes-query-optimizer/LowerToPhysicalWindowedAggregation.cpp: guard the TemporalSequenceAggregationPhysicalFunction include and instantiation with #ifdef NES_ENABLE_MEOS so the translation unit compiles and links when MEOS is disabled.
Add TFLOAT_SIN, TFLOAT_COS, TFLOAT_TAN as per-event scalar operators following the extract-marshaler pattern from W25. Each takes (value:FLOAT64, ts:UINT64), constructs a MEOS single-instant temporal via tfloat_in, applies the trig C function, and extracts the FLOAT64 result via tfloat_start_value. Logical and physical function pairs registered as NES plugins; SQL parser extended with TFLOAT_SIN/COS/TAN tokens in the functionName rule and 2-arg case blocks.
Add TFLOAT_CEIL, TFLOAT_FLOOR, TFLOAT_RADIANS, TFLOAT_DEGREES as per-event scalar operators following the extract-marshaler pattern from W44. Each takes (value:FLOAT64, ts:UINT64), constructs a MEOS single-instant temporal via tfloat_in, applies the scalar C function, and extracts the FLOAT64 result via tfloat_start_value. TFLOAT_DEGREES hardcodes normalize=false. Logical and physical function pairs registered as NES plugins; SQL parser extended with TFLOAT_CEIL/FLOOR/RADIANS/DEGREES tokens in the functionName rule and 2-arg case blocks.
Add TFLOAT_SHIFT_VALUE and TFLOAT_SCALE_VALUE as per-event scalar operators following the extract-marshaler pattern. Each takes (value:FLOAT64, ts:UINT64, param:FLOAT64), constructs a MEOS single-instant temporal via tfloat_in, applies the shift or scale, and extracts the FLOAT64 result via tfloat_start_value. Logical and physical function pairs registered as NES plugins; SQL parser extended with TFLOAT_SHIFT_VALUE/SCALE_VALUE tokens in the functionName rule and 3-arg case blocks.
…s (W47) Adds per-event NES operators backed by `tnumber_abs` (2-arg: value, ts) and `tfloat_shift_scale_value` (4-arg: value, ts, shift, width). Each operator constructs a single-instant tfloat, applies the MEOS function, and returns the resulting double. Logical/physical function pairs, plugin registrations, grammar tokens, and parser case blocks are included.
Adds per-event NES operators backed by `add_tfloat_float`, `sub_tfloat_float`, `mul_tfloat_float`, and `div_tfloat_float` (3-arg each: value, ts, scalar). Each operator constructs a single-instant tfloat, applies the MEOS arithmetic function against a constant scalar, and returns the resulting double. Logical/physical function pairs, plugin registrations, grammar tokens, and parser case blocks are included.
Adds per-event NES operators backed by `add_tnumber_tnumber`, `sub_tnumber_tnumber`, `mul_tnumber_tnumber`, and `div_tnumber_tnumber` (3-arg each: value1, value2, shared-ts). Each operator constructs two co-instant tfloats, applies the MEOS element-wise arithmetic function, and returns the resulting double. Logical/physical function pairs, plugin registrations, grammar tokens, and parser case blocks are included.
…MBER_TNUMBER NES operators (W50) Adds per-event NES operators backed by temporal_round (3-arg: value, ts, maxdd), tdistance_tfloat_float (3-arg: value, ts, d), and tdistance_tnumber_tnumber (3-arg: value1, value2, ts). Each operator constructs one or two co-instant tfloats, applies the MEOS function, and returns the resulting double. Logical/physical function pairs, plugin registrations, grammar tokens, and parser case blocks are included.
Adds per-event NES operators backed by add_tint_int, sub_tint_int, mul_tint_int, and div_tint_int (3-arg each: value, ts, scalar_int). Each operator constructs a single-instant tint, applies the MEOS integer arithmetic function, and returns the resulting integer as FLOAT64. Logical/physical function pairs, plugin registrations, grammar tokens, and parser case blocks are included.
…E, TINT_SHIFT_SCALE_VALUE NES operators (W52)
…, TBIGINT_TO_TFLOAT, TINT_TO_TBIGINT, TFLOAT_TO_TBIGINT (W54)
…FT_SCALE_VALUE NES operators (W56)
…W57) Adds per-event scalar NES operators for reversed-order float arithmetic on tfloat temporals: ADD_FLOAT_TFLOAT, SUB_FLOAT_TFLOAT, MUL_FLOAT_TFLOAT, DIV_FLOAT_TFLOAT. Each operator takes (scalar:float, value:float, ts:uint64) and delegates to the corresponding MEOS add_float_tfloat / sub_float_tfloat / mul_float_tfloat / div_float_tfloat kernel.
Adds per-event scalar NES operators for reversed-order integer arithmetic on tint temporals: ADD_INT_TINT, SUB_INT_TINT, MUL_INT_TINT, DIV_INT_TINT. Each operator takes (scalar:int, value:int, ts:uint64) and delegates to the corresponding MEOS add_int_tint / sub_int_tint / mul_int_tint / div_int_tint kernel.
… (W59) Adds per-event scalar NES operators for reversed-order bigint arithmetic on tbigint temporals: ADD_BIGINT_TBIGINT, SUB_BIGINT_TBIGINT, MUL_BIGINT_TBIGINT, DIV_BIGINT_TBIGINT. Each operator takes (scalar:bigint, value:bigint, ts:uint64) and delegates to the corresponding MEOS add_bigint_tbigint / sub_bigint_tbigint / mul_bigint_tbigint / div_bigint_tbigint kernel.
Adds per-event scalar NES operators for exponential and logarithmic functions on tfloat temporals: TFLOAT_EXP, TFLOAT_LN, TFLOAT_LOG10. Each operator takes (value:float, ts:uint64) and delegates to the corresponding MEOS tfloat_exp / tfloat_ln / tfloat_log10 kernel.
Adds six per-event ever-comparison scalar operators for tfloat against a float threshold: EVER_EQ_TFLOAT_FLOAT, EVER_GE_TFLOAT_FLOAT, EVER_GT_TFLOAT_FLOAT, EVER_LE_TFLOAT_FLOAT, EVER_LT_TFLOAT_FLOAT, EVER_NE_TFLOAT_FLOAT. Each operator constructs a single-instant tfloat from (value, ts), calls the corresponding MEOS ever_*_tfloat_float kernel returning int, and emits the result as double. Logical and physical function pairs, CMakeLists plugin registrations, AntlrSQL grammar tokens, and AntlrSQLQueryPlanCreator case blocks are included.
…W62) Adds six per-event always-comparison scalar operators for tfloat against a float threshold: ALWAYS_EQ_TFLOAT_FLOAT, ALWAYS_GE_TFLOAT_FLOAT, ALWAYS_GT_TFLOAT_FLOAT, ALWAYS_LE_TFLOAT_FLOAT, ALWAYS_LT_TFLOAT_FLOAT, ALWAYS_NE_TFLOAT_FLOAT. Each operator constructs a single-instant tfloat from (value, ts), calls the corresponding MEOS always_*_tfloat_float kernel returning int, and emits the result as double. Logical and physical function pairs, CMakeLists plugin registrations, AntlrSQL grammar tokens, and AntlrSQLQueryPlanCreator case blocks are included.
Adds six per-event ever-comparison scalar operators for tint against an int threshold: EVER_EQ_TINT_INT, EVER_GE_TINT_INT, EVER_GT_TINT_INT, EVER_LE_TINT_INT, EVER_LT_TINT_INT, EVER_NE_TINT_INT. Each operator constructs a single-instant tint from (value, ts), calls the corresponding MEOS ever_*_tint_int kernel returning int, and emits the result as double. Logical and physical function pairs, CMakeLists plugin registrations, AntlrSQL grammar tokens, and AntlrSQLQueryPlanCreator case blocks are included.
Adds six per-event always-comparison scalar operators for tint against an int threshold: ALWAYS_EQ_TINT_INT, ALWAYS_GE_TINT_INT, ALWAYS_GT_TINT_INT, ALWAYS_LE_TINT_INT, ALWAYS_LT_TINT_INT, ALWAYS_NE_TINT_INT. Each operator constructs a single-instant tint from (value, ts), calls the corresponding MEOS always_*_tint_int kernel returning int, and emits the result as double. Logical and physical function pairs, CMakeLists plugin registrations, AntlrSQL grammar tokens, and AntlrSQLQueryPlanCreator case blocks are included.
…W65) Adds six per-event ever-comparison scalar operators for tbigint against a bigint threshold: EVER_EQ_TBIGINT_BIGINT, EVER_GE_TBIGINT_BIGINT, EVER_GT_TBIGINT_BIGINT, EVER_LE_TBIGINT_BIGINT, EVER_LT_TBIGINT_BIGINT, EVER_NE_TBIGINT_BIGINT. Each operator constructs a single-instant tbigint from (value, ts), calls the corresponding MEOS ever_*_tbigint_bigint kernel returning int, and emits the result as double. Logical and physical function pairs, CMakeLists plugin registrations, AntlrSQL grammar tokens, and AntlrSQLQueryPlanCreator case blocks are included.
… (W66) Adds six per-event always-comparison scalar operators for tbigint against a bigint threshold: ALWAYS_EQ_TBIGINT_BIGINT, ALWAYS_GE_TBIGINT_BIGINT, ALWAYS_GT_TBIGINT_BIGINT, ALWAYS_LE_TBIGINT_BIGINT, ALWAYS_LT_TBIGINT_BIGINT, ALWAYS_NE_TBIGINT_BIGINT. Each operator constructs a single-instant tbigint from (value, ts), calls the corresponding MEOS always_*_tbigint_bigint kernel returning int, and emits the result as double. Logical and physical function pairs, CMakeLists plugin registrations, AntlrSQL grammar tokens, and AntlrSQLQueryPlanCreator case blocks are included.
…W123) Wires five trgeometry-vs-trgeometry per-event scalar operators using a 10-arg dual-VARCHAR two-instant pattern (ref1_wkt:VARCHAR, x1:FLOAT64, y1:FLOAT64, theta1:FLOAT64, ts1:UINT64, ref2_wkt:VARCHAR, x2:FLOAT64, y2:FLOAT64, theta2:FLOAT64, ts2:UINT64) -> FLOAT64. Each physical function builds two 2D trgeometry instants via geom_in(ref_wkt), pose_make_2d, and trgeoinst_make. Operators: EVER_EQ/ALWAYS_EQ/EVER_NE/ALWAYS_NE_TRGEOMETRY_TRGEOMETRY (return 0.0/1.0) and NAD_TRGEOMETRY_TRGEOMETRY (distance or 0.0 on error).
…124) Wires two tpcpoint-vs-geometry per-event scalar operators using a 3-arg pattern (pt_hexwkb:VARCHAR, ts:UINT64, tgt_wkt:VARCHAR) -> FLOAT64. Each physical function deserializes the pcpoint from hex-WKB via pcpoint_from_hexwkb, wraps it as a tpcpoint instant via tpointcloudinst_make, and evaluates the geometry target built from WKT. Operators: EINTERSECTS_TPCPOINT_GEO (returns 0.0/1.0) and NAD_TPCPOINT_GEO (returns nearest-approach distance).
Wires seven quadbin cell operators across four patterns: QUADBIN_POINT_TO_CELL (lon/lat/res→VARCHAR), QUADBIN_IS_VALID_CELL/GET_RESOLUTION/CELL_AREA (cell→FLOAT64), QUADBIN_CELL_TO_QUADKEY (cell→VARCHAR), QUADBIN_CELL_TO_PARENT (cell/res→VARCHAR), QUADBIN_TILE_TO_CELL (x/y/z→VARCHAR). Cell values are UINT64; VARCHAR results use quadbin_index_to_string or quadbin_cell_to_quadkey serialization.
Wires four tjsonb-vs-jsonb per-event scalar comparison operators using a 3-arg pattern (json_str:VARCHAR, ts:UINT64, target_json:VARCHAR) -> FLOAT64. Each physical function deserializes both JSON strings via jsonb_in, wraps the first as a tjsonb instant via tjsonbinst_make, then evaluates the comparison. Operators: EVER_EQ/ALWAYS_EQ/EVER_NE/ALWAYS_NE_TJSONB_JSONB (return 0.0/1.0).
Wires four tjsonb-vs-tjsonb per-event scalar comparison operators using a 4-arg dual-instant pattern (json1:VARCHAR, ts1:UINT64, json2:VARCHAR, ts2:UINT64) -> FLOAT64. Each physical function deserializes two JSON strings via jsonb_in and wraps them as tjsonb instants via tjsonbinst_make. Operators: EVER_EQ/ALWAYS_EQ/EVER_NE/ALWAYS_NE_TJSONB_TJSONB (return 0.0/1.0).
…(W128) Wires GEOM_BOUNDARY, GEOM_CENTROID, GEOM_CONVEX_HULL, GEO_REVERSE, GEO_POINTS, and GEOM_UNARY_UNION as 1-arg (wkt:VARCHAR) -> VARCHAR operators. Each physical function parses the WKT input via geom_in, calls the corresponding MEOS GSERIALIZED-returning function, then serializes the result with geo_as_text into an arena-allocated buffer.
… WKT (W129) Wires GEOM_DIFFERENCE2D, GEOM_INTERSECTION2D, GEOM_SHORTESTLINE2D, and GEOM_SHORTESTLINE3D as 2-arg (wkt1:VARCHAR, wkt2:VARCHAR) -> VARCHAR operators. Each physical function parses both WKT inputs via geom_in, calls the corresponding MEOS GSERIALIZED-returning function, then serializes the result with geo_as_text into an arena-allocated buffer.
…ing WKT (W130) Wires GEO_SET_SRID (wkt,srid:U64), GEO_TRANSFORM (wkt,srid_to:U64), GEO_ROUND (wkt,maxdd:U64), and GEOM_BUFFER (wkt,size:F64,params:VARCHAR) as VARCHAR-returning operators. Each physical function parses the WKT input via geom_in, applies the transformation, and serializes the result with geo_as_text.
Wires LINE_NUMPOINTS (wkt->FLOAT64), LINE_LOCATE_POINT (wkt1,wkt2->FLOAT64), LINE_INTERPOLATE_POINT (wkt,frac:F64,repeat:U64->VARCHAR), and LINE_SUBSTRING (wkt,from:F64,to:F64->VARCHAR). FLOAT64-returning ops call line_numpoints/ line_locate_point directly; VARCHAR-returning ops parse via geom_in and serialize results with geo_as_text.
Wires GEOM_POINT_MAKE2D/3DZ and GEOG_POINT_MAKE2D/3DZ (srid:U64,x,y[,z]:F64-> VARCHAR), GEOM_TO_GEOG and GEOG_TO_GEOM (wkt:VARCHAR->VARCHAR), and GEOG_CENTROID (wkt:VARCHAR,use_spheroid:U64->VARCHAR). Constructor ops call geompoint_make2d/3dz and geogpoint_make2d/3dz directly; conversion and centroid ops parse via geom_in/geog_in and serialize with geo_as_text.
…ection NES operators (W133) Wires GEO_GEO_N (wkt,n:U64->VARCHAR) for nth component of a geometry collection, LINE_POINT_N (wkt,n:U64->VARCHAR) for nth point of a linestring (1-based), and GEOM_INTERSECTION2D_COLL (wkt1,wkt2->VARCHAR) for collection-typed intersection. All ops parse WKT via geom_in and serialize results with geo_as_text.
…erators (W134) Wires GEO_AS_GEOJSON/HEXEWKB/EWKT (wkt->VARCHAR), GEO_FROM_GEOJSON and GEOM_FROM_HEXEWKB (string->WKT), GEO_TRANSFORM_PIPELINE (wkt,pipeline,srid,dir-> VARCHAR), GEOM_MIN_BOUNDING_CENTER (wkt->VARCHAR center WKT), and GEOM_MIN_BOUNDING_RADIUS (wkt->FLOAT64). GeoJSON buffer is MAX_LEN=16384 to accommodate larger JSON payloads.
Wires GEOM_RELATE_PATTERN(wkt1, wkt2, pattern) as a FLOAT64-returning per-event operator (1.0=match, 0.0=no match) via geom_relate_pattern, completing the core static geometry surface in MobilityNebula.
Wires 14 per-event operators for temporal geometry pairs — ever/always_eq/ne, eintersects/ecovers/edisjoint/econtains/etouches, aintersects/acovers/adisjoint/ acontains/atouches — as 6-arg (lon1,lat1,ts1:U64,lon2,lat2,ts2:U64) → FLOAT64 predicates backed by meos_geo_tgeo_tgeo functions.
Wires EDWITHIN_TGEO_TGEO and ADWITHIN_TGEO_TGEO as 7-arg (lon1,lat1,ts1:U64,lon2,lat2,ts2:U64,dist:F64) → FLOAT64 predicates backed by edwithin_tgeo_tgeo / adwithin_tgeo_tgeo. Completes the tgeo_tgeo predicate family in MobilityNebula.
Wires H3INDEX_EQ, H3INDEX_NE, H3INDEX_LT, H3INDEX_LE, H3INDEX_GT, H3INDEX_GE, and H3INDEX_CMP as 2-arg (a:UINT64, b:UINT64) -> FLOAT64 operators backed by h3index_eq/ne/lt/le/gt/ge (bool -> 0.0/1.0) and h3index_cmp (int -> -1.0/0.0/1.0). New in pin ecosystem-pin-2026-06-18b.
Wires H3INDEX_OUT(cell:UINT64) -> VARCHAR, backed by h3index_out, converting an H3Index cell identifier to its 15-character hex string representation (e.g. '8928308280fffff'). New in ecosystem-pin-2026-06-18b.
Wires H3INDEX_IN(hex:VARSIZED) -> UINT64, backed by h3index_in, parsing a 15-character H3Index hex string (e.g. '8928308280fffff') to its UINT64 cell identifier. New in ecosystem-pin-2026-06-18b.
Wires JSON_ARRAY_LENGTH(js:VARSIZED)->FLOAT64, and JSON_TYPEOF/OBJECT_FIELD_TEXT/ARRAY_ELEMENT_TEXT(...)-> VARCHAR, backed by json_in/json_out + json_array_length, json_typeof, json_object_field_text, json_array_element_text. key args use cstring_to_text. New in ecosystem-pin-2026-06-18b.
Wires 15 JSONB per-event operators: JSONB_ARRAY_LENGTH (->F64), JSONB_EQ/NE/LT/LE/GT/GE/CMP/CONTAINS/CONTAINED/EXISTS (->F64), and JSONB_TO_CSTRING/PRETTY/OBJECT_FIELD_TEXT/ARRAY_ELEMENT_TEXT (->VARCHAR). Backed by jsonb_in/jsonb_out + the corresponding jsonb_* functions. Key and text args use cstring_to_text/text_to_cstring. New in ecosystem-pin-2026-06-18b.
Wires 12 Span operators: INTSPAN_MAKE and FLOATSPAN_MAKE (4×FLOAT64 → VARCHAR serialized span), plus INTSPAN/FLOATSPAN LOWER/UPPER/WIDTH (→ FLOAT64 bounds/width) and LOWER_INC/UPPER_INC (→ FLOAT64 0/1 flags). Spans are passed as VARCHAR strings across NES operator boundaries using intspan_in/intspan_out and floatspan_in/floatspan_out. New in ecosystem-pin-2026-06-18b.
Wires 8 Span predicate operators: CONTAINED_INT/FLOAT_SPAN (scalar in span), CONTAINS_SPAN_INT/FLOAT (span contains scalar), and CONTAINED/CONTAINS_SPAN_SPAN for both intspan and floatspan variants. All predicates return FLOAT64 (1.0/0.0). Spans are passed as VARCHAR strings serialized by the W143 MAKE operators. New in ecosystem-pin-2026-06-18b.
Wires TEXT_UPPER, TEXT_LOWER, and TEXT_INITCAP (VARSIZED -> VARCHAR), backed by text_upper, text_lower, and text_initcap. Input is converted with cstring_to_text; output is extracted with text_to_cstring. New in ecosystem-pin-2026-06-18b.
Wires QUADBIN_EQ/NE/LT/LE/GT/GE (UINT64×UINT64 → FLOAT64, 0/1) and QUADBIN_CMP (UINT64×UINT64 → FLOAT64, -1/0/1), backed by quadbin_eq, quadbin_ne, quadbin_lt, quadbin_le, quadbin_gt, quadbin_ge, and quadbin_cmp. Mirrors the H3INDEX_EQ/CMP pattern from W138.
Wires 12 tquadbin temporal predicate operators returning FLOAT64 (0/1): EVER/ALWAYS_EQ/NE_QUADBIN_TQUADBIN (3-arg: static cell first), EVER/ALWAYS_EQ/NE_TQUADBIN_QUADBIN (3-arg: instant first), and EVER/ALWAYS_EQ/NE_TQUADBIN_TQUADBIN (4-arg: two instants). tquadbin instants are built per-call via tquadbininst_make(Quadbin, TimestampTz). New in ecosystem-pin-2026-06-18e.
estebanzimanyi
force-pushed
the
feat/nebula-codegen-w148-trgeometry-geo-predicates
branch
2 times, most recently
from
10d390b to
fa334cd
Compare
…s (W148) Wires 14 trgeometry/geometry per-event spatial predicates returning FLOAT64 (0/1). trgeometry_geo (6-arg: ref_wkt, x, y, theta, ts, tgt_wkt): EINTERSECTS, AINTERSECTS, EDISJOINT, ADISJOINT, ETOUCHES, ATOUCHES, ECOVERS, ACOVERS; EDWITHIN, ADWITHIN add a trailing dist (7-arg). geo_trgeometry (6-arg: tgt_wkt, ref_wkt, x, y, theta, ts): ECONTAINS, ACONTAINS, ECOVERS, ACOVERS. Each trgeometry instant is built per-call from the reference WKT geometry and a 2D pose via trgeometryinst_make(geom, pose_make_2d(x,y,theta), ts). New in ecosystem-pin-2026-06-22a.
Pin the dependency-image MEOS build to the canonical ecosystem pin ecosystem-pin-2026-06-22a (commit 043d5e723422d6134a9f5868a7007c9eae65730b) so the trgeometry spatial-predicate symbols are present at link time.
estebanzimanyi
force-pushed
the
feat/nebula-codegen-w148-trgeometry-geo-predicates
branch
from
fa334cd to
2e44d8a
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Wires 14 trgeometry/geometry per-event spatial predicates returning FLOAT64 (0/1). trgeometry_geo (6-arg: ref_wkt, x, y, theta, ts, tgt_wkt): EINTERSECTS, AINTERSECTS, EDISJOINT, ADISJOINT, ETOUCHES, ATOUCHES, ECOVERS, ACOVERS; EDWITHIN and ADWITHIN add a trailing dist (7-arg). geo_trgeometry (6-arg: tgt_wkt, ref_wkt, x, y, theta, ts): ECONTAINS, ACONTAINS, ECOVERS, ACOVERS. Each trgeometry instant is built per-call from the reference WKT geometry and a 2D pose via trgeometryinst_make(geom, pose_make_2d(x,y,theta), ts). New in ecosystem-pin-2026-06-22a.