Skip to main content
Glama

Import geodata

import_geodata

Writes what a reference of the room holds into the model: into terrain creates (created names the new elements), into site writes values at the named IfcSite, into alignment an IfcAlignment from LandXML; dry_run previews. Never asks a service; the reference is the kept answer (reference_id from get_geodata, or a terrain the person uploaded; read_reference lists them). into terrain: an IfcGeographicElement in the IfcSite from its heights, built as create_terrain builds one: thinned (delatin/1) up to epsilon_m or the mesh limit per body, cells without a height masked and never filled, map metres into project m through the model georeference (another EPSG point by point over a bridge, named in the report). base_z is required, never assumed. Moddl_TerrainLayers keeps the origin: Dgm* (file, window, CRS, bridge) and from a service Geodata* (service, fetch, licence, links, credit); a credit goes with whatever is built from it. Without OrthogonalHeight in the georeference it is refused: map heights would put the ground metres away from the building. Ask the person for the height of 0.00 above the vertical datum (building permit, survey) and write it with set_project_origin, source user. The report's terrain_created.dgm: service and fetch, cells, masked cells, epsilon reached, each number with its source, and ground: the terrain under the building against the storey nearest z 0, warned past 3 m (a datum error, or a slope). into site: one parcel (feature: its id from the get_geodata rows) of a parcels reference at the IfcSite named by GlobalId: Pset_LandRegistration.LandID the official identifier (none: refused); Moddl_Geodaten the service, fetch, licence, links, credit, reference id, ParcelArea from the outline (m2; assumed over a projection) and ParcelOfficialArea only where the row carries one. The report names each value with source and the value it replaced. New revision, effect report. A call that succeeds is not a job that's done. Double-check the ground against the building: render_view from two sides.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "reference_id": "<reference_id>",
      "into": "<into>"
    },
    "name": "import_geodata"
  }
}

Input Schema

NameRequiredDescriptionDefault
intoYesterrain: a terrain from the heights of a GeoTIFF or XYZ reference; site: a parcel of a parcels reference as values of the site; alignment: an Alignment of a LandXML reference
nameNointo terrain: name of the container or the single element
noteNoYour note, stored with the new revision
siteNointo terrain: the IfcSite by GlobalId or Name, needed where the file has several; into site, required: the GlobalId of the IfcSite written
base_zNointo terrain, required: bottom of the terrain body, world z in m; never assumed
layersNointo terrain: top to bottom; the last reaches down to base_z. None or one: one terrain element
windowNointo terrain: [min_e, min_n, max_e, max_n] in metres of the raster EPSG (not project m, not the model EPSG); the heights inside it. Left out: the whole raster
dry_runNoPreview: the planned effect, nothing written
featureNointo site, required: the parcel, by its id in the get_geodata rows (not its feature index) or its official identifier
stationsNointo alignment: stations run with project metres where the map factor of the georeference moves them by more than 1 mm over the length (assumption stations_follow_project_distance, deviation_m, equations shifted in the report); left out, such a file is refused
alignmentNointo alignment: one Alignment of a LandXML reference placed in the viewer, by its name or #index; left out where the file holds one. Written as create_alignment writes it (line, clothoid, arc, PVI with ParaCurve and CircCurve, StaEquation). Each segment is closed against the end the file states (alignment_import.closure, 1 mm); over it the whole alignment is refused with index and type (spiType other than clothoid, UnsymParaCurve). A Cant of the file is written only with rail_head_distance_m; without it the report (alignment_import.cant) says the cant is there and not written, and cant.call is edit_alignment at the alignment written, with the rows of the file and the field to fill (never a second import) The layouts read back with get_element include alignment; edit_alignment takes the layout keys of create_alignment (positive radii, kind, turn)
epsilon_mNointo terrain: largest deviation from the raster heights, m; left out: thinned to the mesh limit, the value reached an assumption
request_idNoYour id for this call: a repeat answers the first result and writes nothing
reference_idYesThe reference to write from: get_geodata reference_id, or an upload read_reference lists
station_tolerance_mNointo alignment: how far in plan, in m, the end of the chain written may lie from the end of the file (the segments close against their own ends, each within 1 mm; the written chain only follows them, so it drifts where a joint kinks). At least 0.001, the default; it only widens. Above it the whole alignment is refused with the drift in the detail and the call with this field in detail.call (the value yours). Reported per segment (closure_m, chain_drift_m), as written_end_gap_m and as assumption chain_drift_tolerance_user
confirm_height_checkNointo terrain: an upload is held against catalogue terrain of the room (dgm.height_check); a cm/mm, ellipsoidal or gross reading is refused REFERENCE_HEIGHT_CHECK. true writes it anyway, only after the person confirmed the heights
rail_head_distance_mNointo alignment: takes the Cant of the file (CantStation, appliedCant, rotationPoint, clothoid ramps) as the cant layout. The distance of the rail heads the roll is formed with (about 1.5 m on standard gauge): yours to name, never the gauge of the file (about 1.435 m); the report shows both. Source user, between the gauge of the file and twice it (rail_head_distance_implausible: a value in mm or another unit is refused with both values). Refused if the Cant cannot be taken whole (index and type), reaches the distance, or the alignment has no profile; a clothoid ramp is read as linear (assumption cant_clothoid_linear with its deviation_m)

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed3 schema fields changed
    • changedInput schema / properties / alignment / description
      Previous value: -"into alignment: one Alignment of a LandXML reference placed in the viewer, by its name or #index; left out where the file holds one. Written as create_elements alignment writes it (line, clothoid, arc, PVI with ParaCurve and CircCurve, StaEquation). Each segment is closed against the end the file states (alignment_import.closure, 1 mm); over it the whole alignment is refused with index and type (spiType other than clothoid, UnsymParaCurve); Cant is not read. The layouts read back with get_element include alignment; edit_alignment takes the layout keys of create_elements alignment (positive radii, kind, turn)"New value: +"into alignment: one Alignment of a LandXML reference placed in the viewer, by its name or #index; left out where the file holds one. Written as create_alignment writes it (line, clothoid, arc, PVI with ParaCurve and CircCurve, StaEquation). Each segment is closed against the end the file states (alignment_import.closure, 1 mm); over it the whole alignment is refused with index and type (spiType other than clothoid, UnsymParaCurve). A Cant of the file is written only with rail_head_distance_m; without it the report (alignment_import.cant) says the cant is there and not written, and cant.call is edit_alignment at the alignment written, with the rows of the file and the field to fill (never a second import) The layouts read back with get_element include alignment; edit_alignment takes the layout keys of create_alignment (positive radii, kind, turn)"
    • addedInput schema / properties / rail_head_distance_m
      Added value: +{
      +  "description": "into alignment: takes the Cant of the file (CantStation, appliedCant, rotationPoint, clothoid ramps) as the cant layout. The distance of the rail heads the roll is formed with (about 1.5 m on standard gauge): yours to name, never the gauge of the file (about 1.435 m); the report shows both. Source user, between the gauge of the file and twice it (rail_head_distance_implausible: a value in mm or another unit is refused with both values). Refused if the Cant cannot be taken whole (index and type), reaches the distance, or the alignment has no profile; a clothoid ramp is read as linear (assumption cant_clothoid_linear with its deviation_m)",
      +  "exclusiveMinimum": 0,
      +  "type": "number"
      +}
    • addedInput schema / properties / station_tolerance_m
      Added value: +{
      +  "description": "into alignment: how far in plan, in m, the end of the chain written may lie from the end of the file (the segments close against their own ends, each within 1 mm; the written chain only follows them, so it drifts where a joint kinks). At least 0.001, the default; it only widens. Above it the whole alignment is refused with the drift in the detail and the call with this field in detail.call (the value yours). Reported per segment (closure_m, chain_drift_m), as written_end_gap_m and as assumption chain_drift_tolerance_user",
      +  "exclusiveMinimum": 0,
      +  "type": "number"
      +}
  2. Changed4 schema fields changed
    • addedInput schema / properties / alignment
      Added value: +{
      +  "description": "into alignment: one Alignment of a LandXML reference placed in the viewer, by its name or #index; left out where the file holds one. Written as create_elements alignment writes it (line, clothoid, arc, PVI with ParaCurve and CircCurve, StaEquation). Each segment is closed against the end the file states (alignment_import.closure, 1 mm); over it the whole alignment is refused with index and type (spiType other than clothoid, UnsymParaCurve); Cant is not read. The layouts read back with get_element include alignment; edit_alignment takes the layout keys of create_elements alignment (positive radii, kind, turn)",
      +  "minLength": 1,
      +  "type": "string"
      +}
    • changedInput schema / properties / into / description
      Previous value: -"terrain: a terrain from the heights of a GeoTIFF or XYZ reference; site: a parcel of a parcels reference as values of the site"New value: +"terrain: a terrain from the heights of a GeoTIFF or XYZ reference; site: a parcel of a parcels reference as values of the site; alignment: an Alignment of a LandXML reference"
    • changedInput schema / properties / into / enum
      Previous value: -[
      -  "terrain",
      -  "site"
      -]New value: +[
      +  "terrain",
      +  "site",
      +  "alignment"
      +]
    • addedInput schema / properties / stations
      Added value: +{
      +  "description": "into alignment: stations run with project metres where the map factor of the georeference moves them by more than 1 mm over the length (assumption stations_follow_project_distance, deviation_m, equations shifted in the report); left out, such a file is refused",
      +  "enum": [
      +    "project_distance"
      +  ],
      +  "type": "string"
      +}
  3. Added

TDQS

A4.1/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already declare readOnlyHint=false and destructiveHint=false, and the description goes well beyond them: it states that nothing is fetched live, that a request_id repeat answers the first result and writes nothing, that a new revision and effect report are produced, that a credit travels with service-derived data, and that terrain is thinned within epsilon_m or the mesh limit with masked cells never filled. It also spells out hard refusal cases (missing OrthogonalHeight, height-check failures, closure beyond 1 mm). The volume is high, but the disclosure is genuine rather than restating annotations.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness2/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is a dense wall of run-on clauses with semicolon-chained caveats; key facts such as the three 'into' branches and the required base_z are buried mid-paragraph, and several sentences restate the schema verbatim. It is not front-loaded and is oversized relative to what can be usefully read before a call.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 17-parameter, three-mode write tool with no output schema, it covers the prerequisites (georeference OrthogonalHeight, site GlobalId), the refusal conditions per mode, and the structure of the returned report (terrain_created.dgm, ground distance to the storey nearest z 0, the replaced values, assumption names). Almost nothing needed to call it correctly is left unstated.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100% and the schema already explains most fields, so the baseline is 3; the description nevertheless adds meaning the schema lacks: epsilon_m's fallback to the mesh limit and delatin/1 thinning, window being in raster EPSG rather than project metres, feature being the get_geodata row id rather than a feature index, and rail_head_distance_m' being the rail-head distance rather than the file's gauge. Some passages merely repeat the schema (base_z never assumed, confirm_height_check semantics), which keeps it below 5.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description states a specific verb and resource ('Writes what a reference of the room holds into the model') and enumerates three distinct write targets: terrain (IfcGeographicElement in the IfcSite), site (parcel values on the named IfcSite), and alignment (IfcAlignment from LandXML). It clearly distinguishes itself from get_geodata, create_terrain, create_alignment and edit_alignment, but the opening clause is wordy and the actual effect is only fully clear after parsing several run-on sentences.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

It names the alternatives and the conditions that select them explicitly: reference_id comes from get_geodata or from read_reference, a missing OrthogonalHeight must be resolved with set_project_origin rather than imported, a cant must be finished with edit_alignment rather than a second import, and the result should be verified with render_view from two sides. When-not-use is covered through the refusal conditions (REFERENCE_HEIGHT_CHECK, missing LandID, no rail_head_distance_m).

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources