mapsmith 0.8.0__tar.gz → 0.9.0__tar.gz

{mapsmith-0.8.0 → mapsmith-0.9.0}/.zenodo.json

@@ -1,7 +1,7 @@
 {
   "upload_type": "software",
   "title": "MapSmith: geospatial processing over MCP, with verifiable provenance",
-  "version": "0.8.0",
+  "version": "0.9.0",
   "creators": [
     {"name": "The MapSmith project"}
   ],

{mapsmith-0.8.0 → mapsmith-0.9.0}/CHANGELOG.md

@@ -6,6 +6,110 @@ All notable changes to MapSmith are documented here, in the format of
 
 ## [Unreleased]
 
+## [0.9.0] - 2026-10-08
+
+Records still declare `spec_version` `1.0.0-draft.10`, and no key is renamed.
+Three keys of MapSmith's own are new, each declared with the reason it is not
+a synonym of anything the specification has: `x-mapsmith:areas_measured_on` in
+`crs_decisions`, from `apportion_by_area`, and `x-mapsmith:ballpark_share` and
+`x-mapsmith:outside_area_of_use` inside `crs_decisions.transformation`. Three
+check names are new, all from `apportion_by_area`:
+`x-mapsmith:extensive_total_preserved`, `x-mapsmith:every_source_value_placed`
+and `x-mapsmith:no_zone_gives_more_than_it_holds`. **What
+`reproject_layer` writes changes in two cases**, both described under *Fixed*:
+coordinates outside every published operation's area are no longer moved by
+an operation from elsewhere, and a geographic CRS counted from another
+meridian whose operations are published only for its Greenwich twin
+(EPSG:4806, Monte Mario from Rome) goes through that twin. Shapely 2.1 and
+pyarrow 23.0.1 or later are required.
+
+### Added
+
+- **`apportion_by_area`: move counts and rates from the zones they are reported
+  in to zones that cross them.** Cases by municipality to health districts,
+  census-tract population to school catchments. Each column is declared
+  extensive (a count, split by share of area) or intensive (a rate, averaged by
+  area); a call declaring neither is refused, because splitting a rate or
+  averaging a count gives a plausible wrong number. Each count comes with
+  `_min` and `_max` that hold however it is distributed inside the source
+  zones; rate bounds rest on the same uniformity the estimate assumes, and the
+  record says which is which. Zones are cut in the source zones' own CRS,
+  where a boundary both layers share is the same line, and a zone counts as
+  wholly inside a target by a topological test, not by a ratio of areas.
+  Shares are ratios of areas measured on an equal-area surface -- the source
+  CRS when it is one, otherwise a plane centred on the zones on their own
+  datum, named in the record -- so the pieces of each zone add up to it.
+  Whether targets overlap is decided zone by zone, so a few square metres of
+  overlapping boundaries are said in the record where they put a zone's value
+  in two targets, and a check fails if a zone with no overlap gives out more
+  than it holds. Invalid zones are repaired and the repair recorded; a zone
+  the repair leaves with no area or with a coordinate that is not a number,
+  zones spread over more than a hemisphere, layers writing longitude in two
+  conventions, and a target that the source CRS would tear at one of its
+  seams (the 180th meridian, a pole: caught by area and by position, whatever
+  the units) are refused; targets keep their own CRS; and the counts written
+  are checked against the pieces they came from.
+
+### Fixed
+
+- **`crs_decisions.transformation` now describes what PROJ did to the data's
+  own coordinates.** Which operation PROJ selects depends on where a
+  coordinate is, and the record asked at one point chosen from the source CRS
+  alone -- for EPSG:4326, the middle of the world, in the Gulf of Guinea. It
+  was wrong in both directions: WGS 84 data in Europe moved to EPSG:3035, or in
+  England to EPSG:27700, was recorded as a ballpark (`is_ballpark: true`) where
+  PROJ applied a published 1 m or 2 m shift; and NAD27 coordinates in Italy were
+  recorded as a 7 m shift where PROJ applied none -- the error a ballpark record
+  exists to report, hidden by it. Every operation that moves an input now hands
+  over the moved layer, PROJ is asked what it did at a spread of the data's own
+  points, inside and outside the areas of use of the operations it has -- outside
+  them it applies no shift for some pairs (NAD27) and stretches a published
+  operation for others (DHDN, the Swiss grid), so the answer is PROJ's and not
+  inferred. A layer straddling the edge of a datum grid is recorded as a
+  ballpark with `x-mapsmith:ballpark_share` saying how many of its vertices got
+  no shift; `x-mapsmith:outside_area_of_use` says how many lie outside the area
+  of every operation published for the pair, whatever PROJ then did with them;
+  a move of no coordinates at all records that nothing was transformed; and a
+  check with no data to hand asks where the two CRSs' areas of use overlap.
+- **`reproject_layer` no longer applies an operation outside its area of use,
+  and its output changes where it did.** When PROJ's own choice at its probe
+  point was a ballpark it substituted the first published operation for the
+  pair, wherever the data was: WGS 84 data in Madrid moved into EPSG:3003 by
+  the Italian operation came out 84.5 m from PROJ's own result, and WGS 84
+  data in Italy moved into EPSG:27700 by the British one, 248 m. Now it
+  substitutes only an operation whose area of use contains every vertex, and
+  otherwise leaves each point where PROJ does. A geographic CRS counted from
+  another meridian whose operations are published only for its Greenwich twin
+  goes through that twin -- EPSG:4806 (Monte Mario, Rome) and EPSG:4819 (Nord
+  Sahara 1959, Paris) among the 23 tried; NTF Paris, EPSG:4807, has an
+  operation of its own and does not change. A station in Piedmont on EPSG:4806
+  now lands on the published 4 m operation's answer, where it was 6.3 m off.
+- **`better_available_m` names only an operation that covers the data and is
+  not installed.** It reported any published operation for the pair: EPSG:4806
+  data in Piedmont was told a 44 m operation needed its grid installed, when
+  that operation needs none and is valid in Sardinia only.
+- **A record written on a failing path no longer claims an earlier run's
+  output.** With a dataset already at the output path, a run refused by its
+  preconditions, or one that crashed before writing, recorded the digest of
+  the earlier bytes as its own output -- and a lineage walk by digest then
+  attributed them to the inputs of the run that failed -- and its record
+  replaced theirs, leaving them with none. Such a record now omits `output`
+  and says why, and where a record on disk still describes the bytes at the
+  path it is left in place and the failing run's record is written beside it
+  as `<output>.failed.provenance.json`, which the error names. Bytes the
+  failing run did write keep their digest, including when a check of the
+  output crashes after the write.
+
+### Changed
+
+- **Shapely 2.1 or later is required** (was 2.0): `apportion_by_area` uses its
+  maximum inscribed circle to check that a reprojected target still holds its
+  own interior.
+- **pyarrow 23.0.1 or later is required** (was 21), the release fixing
+  CVE-2026-25087, a use-after-free reading Arrow IPC files; MapSmith reads none,
+  and the floor keeps an install at the minimum from carrying it. pytest 9.0.3
+  or later for the test extra.
+
 ## [0.8.0] - 2026-10-05
 
 Records declare `spec_version` `1.0.0-draft.10`. `draft.9` defines the digest

{mapsmith-0.8.0 → mapsmith-0.9.0}/CITATION.cff

@@ -12,8 +12,8 @@ identifiers:
     value: "10.5281/zenodo.22213091"
     description: "Concept DOI — always resolves to the latest release"
 license: AGPL-3.0-or-later
-version: "0.8.0"
-date-released: "2026-10-05"
+version: "0.9.0"
+date-released: "2026-10-08"
 keywords:
   - GIS
   - geospatial

{mapsmith-0.8.0 → mapsmith-0.9.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: mapsmith
-Version: 0.8.0
+Version: 0.9.0
 Summary: Professional-grade GIS geoprocessing over MCP, with verifiable provenance
 Project-URL: Homepage, https://github.com/mapsmith-ai/MapSmith
 Project-URL: Repository, https://github.com/mapsmith-ai/MapSmith
@@ -21,10 +21,10 @@ Requires-Dist: duckdb<2,>=1.5
 Requires-Dist: geopandas>=1.1.2
 Requires-Dist: mcp<2,>=1.28.1
 Requires-Dist: model2vec<1,>=0.9
-Requires-Dist: pyarrow>=21
+Requires-Dist: pyarrow>=23.0.1
 Requires-Dist: pyogrio>=0.9
 Requires-Dist: pyproj>=3.6
-Requires-Dist: shapely>=2.0
+Requires-Dist: shapely>=2.1
 Provides-Extra: postgres
 Requires-Dist: psycopg[binary]>=3.2; extra == 'postgres'
 Provides-Extra: raster
@@ -35,7 +35,7 @@ Provides-Extra: sedona
 Requires-Dist: apache-sedona[db]>=1.9; extra == 'sedona'
 Provides-Extra: test
 Requires-Dist: jsonschema>=4.23; extra == 'test'
-Requires-Dist: pytest>=8.0; extra == 'test'
+Requires-Dist: pytest>=9.0.3; extra == 'test'
 Requires-Dist: ruff<0.17,>=0.16; extra == 'test'
 Provides-Extra: whitebox
 Requires-Dist: whitebox-workflows<3,>=2.0.6; extra == 'whitebox'
@@ -49,7 +49,7 @@ Description-Content-Type: text/markdown
 [![PyPI](https://img.shields.io/pypi/v/mapsmith)](https://pypi.org/project/mapsmith/)
 [![Container](https://img.shields.io/badge/ghcr.io-mapsmith--ai%2Fmapsmith-2496ED?logo=docker&logoColor=white)](https://github.com/mapsmith-ai/MapSmith/pkgs/container/mapsmith)
 [![MCP](https://img.shields.io/badge/Model_Context_Protocol-server-654FF0)](https://modelcontextprotocol.io)
-[![License: AGPL-3.0](https://img.shields.io/badge/license-AGPL--3.0-blue)](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/LICENSE)
+[![License: AGPL-3.0](https://img.shields.io/badge/license-AGPL--3.0-blue)](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/LICENSE)
 
 **Ask a question about your data. Get an analysis — and the record that proves it.**
 
@@ -88,12 +88,12 @@ MapSmith's private output: JSON Schema, a toolchain-free validator, a conformanc
 hundred-line emitter that never imports MapSmith. Records carry `spec_version`, and CI validates
 real MapSmith output against the spec's own validator. The specification is archived and citable
 as [10.5281/zenodo.22205213](https://doi.org/10.5281/zenodo.22205213), a concept DOI that
-resolves to the latest *archived* draft. Records written by 0.8.0 declare `1.0.0-draft.10`, and
+resolves to the latest *archived* draft. Records written by 0.8.0 and 0.9.0 declare `1.0.0-draft.10`, and
 0.7.x's declare `1.0.0-draft.8`; if the DOI shows an earlier one, the archive has not caught up, and the
 [repository](https://github.com/mapsmith-ai/manifest-spec) holds the text they are written against. The names MapSmith puts
 in a record beyond the ones the specification defines — every extension check name, and every
 extension field in `crs_decisions` with the reason it is not a synonym of a key the specification
-already has — are listed in [`docs/manifest-vocabulary.md`](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/docs/manifest-vocabulary.md),
+already has — are listed in [`docs/manifest-vocabulary.md`](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/docs/manifest-vocabulary.md),
 generated from the source rather than maintained by hand.
 
 Evidence before promises. A correctness suite in its own organisation,
@@ -102,9 +102,9 @@ computed on paper before any system runs: **0.00 silent errors, nothing skipped*
 0.9032 for the obvious way of writing the same code. The runs from 2026-08-30 to 2026-09-15 said
 0.00 too, and on one trap they were wrong: its truth was wrong in MapSmith's favour for 26 days, the error was ours
 twice over, and the [erratum](https://github.com/argleton/argleton/blob/main/results/README.md#erratum-2026-09-25-trap-024) says so. Every run since 2026-09-26, the current one
-included, is scored against the corrected truth. Getting here cost seven defects the suite sent back, and they are listed. Alongside it: an [A/B on GABench](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/docs/benchmarks.md) whose
+included, is scored against the corrected truth. Getting here cost seven defects the suite sent back, and they are listed. Alongside it: an [A/B on GABench](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/docs/benchmarks.md) whose
 headline is a null result — with the analysis that took our own positive number apart —
-[notebooks](https://github.com/mapsmith-ai/MapSmith/tree/v0.8.0/examples/) on a real USGS DEM of Mount St. Helens, an
+[notebooks](https://github.com/mapsmith-ai/MapSmith/tree/v0.9.0/examples/) on a real USGS DEM of Mount St. Helens, an
 [in-chat map panel](#see-results-inside-the-chat) that shows the verification status of
 every layer it draws, and a
 [measurement of our own tool discovery](#finding-the-right-operation) that retracted two numbers
@@ -113,7 +113,7 @@ this page had already published — including the one in the bullet list below.
 ## A whole analysis, start to finish
 
 This is the thing MapSmith is for, and it is not a single tool call. One question — six
-parcels, a river, an elevation grid — becomes a plan of five operations chosen out of 77,
+parcels, a river, an elevation grid — becomes a plan of five operations chosen out of 78,
 validated before anything runs, executed step by step, and recorded: the search that
 narrowed the catalogue, the arguments that mattered, the CRS decision behind each metric
 step, and the checks that ran on every result.
@@ -124,7 +124,7 @@ plan, reads the manifests, and writes what follows; `tests/test_worked_example.p
 page and that script disagree. The position column is BM25's rather than the default engine's,
 because a published figure should not depend on whether a model download succeeded on the machine
 that built the page — the narrowing, which is the point, is identical on both. Two things worth watching: the middle column, where the catalogue
-goes from 77 operations to a handful the caller can read; and the CRS column, where every
+goes from 78 operations to a handful the caller can read; and the CRS column, where every
 metric operation says which coordinate system it moved the data into and why.
 
 <!-- worked-example:start -->
@@ -135,15 +135,15 @@ flowchart TB
   ASK --> PLAN{{"plan validated<br/>before anything runs"}}
   PLAN -. "rejected: FORWARD_REFERENCE" .-> BAD["'mask_path' references '$buffer' which runs later — move step 'buffer' before 'near'"]
   BAD:::bad
-  BUFFER["<b>buffer_layer</b><br/>77 operations &rarr; 27 candidates &rarr; chosen<br/>CRS EPSG:32610<br/>9/9 checks"]
+  BUFFER["<b>buffer_layer</b><br/>78 operations &rarr; 27 candidates &rarr; chosen<br/>CRS EPSG:32610<br/>9/9 checks"]
   PLAN --> BUFFER
-  NEAR["<b>clip_layer</b><br/>77 operations &rarr; 12 candidates &rarr; chosen<br/>CRS EPSG:4326<br/>12/12 checks"]
+  NEAR["<b>clip_layer</b><br/>78 operations &rarr; 13 candidates &rarr; chosen<br/>CRS EPSG:4326<br/>12/12 checks"]
   BUFFER --> NEAR
-  HEIGHT["<b>zonal_statistics</b><br/>77 operations &rarr; 2 candidates &rarr; chosen<br/>CRS EPSG:4326<br/>7/7 checks"]
+  HEIGHT["<b>zonal_statistics</b><br/>78 operations &rarr; 2 candidates &rarr; chosen<br/>CRS EPSG:4326<br/>7/7 checks"]
   NEAR --> HEIGHT
-  AREA["<b>measure_area</b><br/>77 operations &rarr; 24 candidates &rarr; chosen<br/>CRS WGS 84 &#40;ellipsoidal&#41;<br/>10/10 checks"]
+  AREA["<b>measure_area</b><br/>78 operations &rarr; 24 candidates &rarr; chosen<br/>CRS WGS 84 &#40;ellipsoidal&#41;<br/>10/10 checks"]
   HEIGHT --> AREA
-  FILTER["<b>select_features</b><br/>77 operations &rarr; 24 candidates &rarr; chosen<br/>CRS EPSG:4326<br/>10/10 checks"]
+  FILTER["<b>select_features</b><br/>78 operations &rarr; 24 candidates &rarr; chosen<br/>CRS EPSG:4326<br/>10/10 checks"]
   AREA --> FILTER
   OUT[["3 parcels, each with elevation and ground area"]]
   FILTER --> OUT
@@ -152,11 +152,11 @@ flowchart TB
 
 | what the agent asks for | it declares | candidates | picked | at position |
 |---|---|---|---|---|
-| “everything within one and a half kilometres of the river” | vector, dataset:vector, 1 dataset(s), line | **27** of 77 | `buffer_layer` | 2 |
-| “keep only the parcels that fall inside that strip” | vector, dataset:vector, 2 dataset(s), polygon | **12** of 77 | `clip_layer` | 1 |
-| “how high is the ground under each of these parcels” | raster, dataset:vector, 2 dataset(s), polygon | **2** of 77 | `zonal_statistics` | 2 |
-| “how big is each one on the ground” | vector, dataset:vector, 1 dataset(s), polygon | **24** of 77 | `measure_area` | 1 |
-| “drop the ones where the ground is above 120 metres” | vector, dataset:vector, 1 dataset(s), polygon | **24** of 77 | `select_features` | 2 |
+| “everything within one and a half kilometres of the river” | vector, dataset:vector, 1 dataset(s), line | **27** of 78 | `buffer_layer` | 2 |
+| “keep only the parcels that fall inside that strip” | vector, dataset:vector, 2 dataset(s), polygon | **13** of 78 | `clip_layer` | 1 |
+| “how high is the ground under each of these parcels” | raster, dataset:vector, 2 dataset(s), polygon | **2** of 78 | `zonal_statistics` | 2 |
+| “how big is each one on the ground” | vector, dataset:vector, 1 dataset(s), polygon | **24** of 78 | `measure_area` | 1 |
+| “drop the ones where the ground is above 120 metres” | vector, dataset:vector, 1 dataset(s), polygon | **24** of 78 | `select_features` | 2 |
 
 | step | operation | arguments that mattered | CRS decision, recorded | checks |
 |---|---|---|---|---|
@@ -266,17 +266,31 @@ or from a terminal: `code --add-mcp '{"name":"mapsmith","command":"uvx","args":[
 To check it runs before wiring a client, `uvx mapsmith` starts the server on stdio
 (Ctrl-C to quit) — it speaks MCP, not a CLI, so a silent prompt means it is working.
 
-This page describes **0.8.0**, which is what that command installs. When `main` runs ahead of
+This page describes **0.9.0**, which is what that command installs. When `main` runs ahead of
 the published artifact this paragraph says so and names the difference — a reader should never
-have to find out by calling a tool that is not there. **If you run 0.4.0 to 0.6.1, 0.7.0 and
+have to find out by calling a tool that is not there. **From 0.8.0, 0.9.0** adds
+`apportion_by_area`, which moves counts and rates from the zones they are reported in to zones
+that cross them, and makes a record's `transformation` describe what PROJ did to the data's own
+coordinates rather than to a point chosen from the CRS. **That changes what `reproject_layer`
+writes in two cases**: coordinates outside every published operation's area are no longer
+moved by an operation from elsewhere, and a datum counted from another meridian whose
+operations are published only for its Greenwich twin -- Monte Mario from Rome, EPSG:4806 --
+goes through that twin. The [changelog](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/CHANGELOG.md#090---2026-10-08) names both. **If you run 0.4.0 to 0.6.1, 0.7.0 and
 later change answers you have**: every position read from a point-registered raster
 (`AREA_OR_POINT=Point`, the USGS and Copernicus DEMs) was half a cell off, and the
-[changelog](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/CHANGELOG.md) says which operations and why. **From 0.7.x, 0.8.0** adds
+[changelog](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/CHANGELOG.md) says which operations and why. **From 0.7.x, 0.8.0** adds
 `cluster_points_by_distance` and a `geometry` facet for the catalogue search, digests a
 shapefile with its attributes and its CRS, reads any single-layer file geodatabase, and fixes
 operations that accepted the wrong geometry and answered anyway (`zonal_statistics` on line
 zones returned their length as a pixel count).
 
+**If you read manifests, 0.9.0 renames no key**: records still declare `1.0.0-draft.10`.
+Three keys of MapSmith's own are new: `crs_decisions.x-mapsmith:areas_measured_on`, from
+`apportion_by_area`, and `x-mapsmith:ballpark_share` and `x-mapsmith:outside_area_of_use`
+inside `transformation`; so are three check names, all from `apportion_by_area`. A run that
+fails where an earlier run's dataset still sits writes its record beside the earlier one, as
+`<output>.failed.provenance.json`, instead of replacing it.
+
 **If you read manifests, 0.8.0 renames one key**: records declare `1.0.0-draft.10`, and a
 second input's own georeferencing facts, `inputs[].x-mapsmith:environment` on 0.7.x, are
 `inputs[].environment`, the name the specification gives them since `draft.10`. Each input of
@@ -316,7 +330,7 @@ for it:
 ```json
 {
   "spec_version": "1.0.0-draft.10",
-  "producer": {"name": "mapsmith", "version": "0.8.0"},
+  "producer": {"name": "mapsmith", "version": "0.9.0"},
   "operation": "buffer_layer",
   "parameters": {"distance_meters": 300.0},
   "inputs": [{
@@ -397,7 +411,7 @@ For the **analysis** rather than the step, `run_operation` with `get_lineage` ta
   without the LLM is in there. No AI slop.
 - **The engines compute, the model orchestrates.** Geometry and numbers only ever come
   from deterministic tool executions — never from model output.
-- **77 operations, reached through 28 goal-level tools.** The catalogue is the breadth and
+- **78 operations, reached through 28 goal-level tools.** The catalogue is the breadth and
   the tool list is the choice; it is searchable rather than dumped, because tool-selection
   accuracy degrades once
   a few dozen tools are exposed at once, and fastest when two of them apply to the same
@@ -408,7 +422,7 @@ For the **analysis** rather than the step, `run_operation` with `get_lineage` ta
 - **Model-agnostic infrastructure.** Claude, GPT, Qwen, Kimi, GLM — anything that speaks
   MCP, cloud or local. The leverage is better contracts (typed plans, actionable error
   codes, a searchable catalog), not weights we would have to maintain. See
-  [the manifesto](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/MANIFESTO.md).
+  [the manifesto](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/MANIFESTO.md).
 
 ## Tools
 
@@ -446,8 +460,8 @@ For the **analysis** rather than the step, `run_operation` with `get_lineage` ta
 ### Finding the right operation
 
 Those are the tools an agent chooses between. Behind them the **catalog** holds every
-operation MapSmith can perform — 77 today, and 52 of them have no tool of their own — and
-it is built to hold thousands. (Two of the 77 are marked `planned` and say so when asked:
+operation MapSmith can perform — 78 today, and 53 of them have no tool of their own — and
+it is built to hold thousands. (Two of the 78 are marked `planned` and say so when asked:
 the roadmap is in the catalog on purpose, so an agent can answer "not yet" instead of
 inventing a call.)
 
@@ -466,11 +480,11 @@ was shown this catalog, because a model handed the entry writes a paraphrase of
 
 | what the caller declares | candidates left | BM25, found@3 | embeddings, found@3 | **right answer in what comes back** |
 |---|---|---|---|---|
-| nothing — words alone | 77 | 29% | 19% | 29% |
-| what data I have | 51 | 35% | 22% | 36% |
-| + what I want back | 32 | 45% | 40% | 54% |
-| + how many datasets I have | 17 | 54% | 47% | 84% |
-| **+ what geometry it holds** | **15** | **61%** | **58%** | **98%** |
+| nothing — words alone | 78 | 29% | 18% | 29% |
+| what data I have | 52 | 34% | 21% | 35% |
+| + what I want back | 32 | 44% | 38% | 53% |
+| + how many datasets I have | 17 | 53% | 45% | 84% |
+| **+ what geometry it holds** | **15** | **60%** | **56%** | **98%** |
 
 **The last row is new on 2026-09-29, and the row above it shows why.** Thirty-one operations
 now take one vector layer to another, one more than the threshold below, so a caller who
@@ -493,7 +507,7 @@ published under a sentence promising it could be checked.
 The two also differ in a way worth seeing, and this page had it backwards
 until 2026-08-30. It said the embedding engine overtakes BM25 once the facets
 have narrowed. It does not overtake it anywhere: BM25 leads at every row of the
-table above, by ten points on words alone and by three at the fullest
+table above, by eleven points on words alone and by four at the fullest
 declaration, where the entries that survive are told apart by the words that
 distinguish them — which is what `distinguishes` is for, and what an exact term
 either matches or does not. The embedding engine earns its place on the
@@ -505,7 +519,7 @@ only the table beside it.)*
 
 The last column is not an accuracy figure — it is a property, and the 98% rather than 100% is
 worth a sentence. The narrowing never drops the right operation: that is asserted per entry and
-holds for all 77. What the column measures is whether the surviving set was small enough to hand
+holds for all 78. What the column measures is whether the surviving set was small enough to hand
 over WHOLE, and for a handful of requests it still is not, so those fall back to a ranked
 shortlist and the answer can be outside the top three. Ranking decides the order; it does not
 decide membership; and the 2% is the gap between "cannot lose the answer" and "can show you all
@@ -544,18 +558,18 @@ of comparison this page exists to refuse.)
 That is the shape of the trade, and it says when the next facet is due. The figure to watch is
 not found@3 — a ranker will always get worse as the catalogue grows, and it is a hint. It is the
 **average** surviving set at the fullest declaration, the fourth column of that table:
-9 at 51 operations, 14 at 61, 16 at 72, 16 at 74, 16 at 75, 17 at 76, 15 at 77. (The last point
-is the first with the geometry declared; without it the set would be 17.) When that crosses the threshold, delivery stops being a
+9 at 51 operations, 14 at 61, 16 at 72, 16 at 74, 16 at 75, 17 at 76, 15 at 77, 15 at 78. (The
+point at 77 is the first with the geometry declared; without it the set would be 17.) When that crosses the threshold, delivery stops being a
 property and starts being a ranking again, and the lasting answer is another fact the caller
 already knows. A bigger threshold is an answer only as far as the caller still chooses well
 among more, and that is measured below, not assumed. (It said *median* until 2026-08-29, and published the mean:
-the median at 77 is 12. The distribution is skewed — most requests leave a small set and a few
+the median at 78 is 13. The distribution is skewed — most requests leave a small set and a few
 leave a large one — so the two numbers say different things and the mean is the pessimistic
 one, which is the right one to watch.)
 
 The requests, both labels and the harness are all in the repository:
-[`tests/data/discovery_queries.json`](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/tests/data/discovery_queries.json) and
-[`benchmarks/discovery_report.py`](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/benchmarks/discovery_report.py), which recomputes every
+[`tests/data/discovery_queries.json`](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/tests/data/discovery_queries.json) and
+[`benchmarks/discovery_report.py`](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/benchmarks/discovery_report.py), which recomputes every
 number above from those files with no network and no model — so they can be checked rather than
 believed, and `tests/test_discovery_report.py` fails if this page and the harness disagree. The
 one exception is the 69%: reproducing that needs the model that did the choosing, and the report
@@ -565,15 +579,15 @@ choices are in the repository, so its figures can be recomputed without the mode
 **So it hands over the set instead of picking for you.** At thirty survivors or fewer `list_operations`
 answers with `status: "choose"`: every candidate, ordered as a hint that says it is a hint, each
 carrying the sentence that separates it from its neighbours. Over those 118 requests the surviving
-set has a median of 12 and exceeds 30 for two of them — which is the 98% in the table above,
+set has a median of 13 and exceeds 30 for two of them — which is the 98% in the table above,
 seen from the other side. The payload is about 2,100 tokens, less than one wrong operation costs
 to run and undo.
 
 **Thirty is measured, and a larger threshold was tried first.** Raising it was the obvious answer
 to the thirty-first vector-to-vector operation, and models get better at reading long lists, so
 before moving it we measured whether one still chooses as well from more
-([`benchmarks/choice_at_scale.py`](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/benchmarks/choice_at_scale.py), every choice in
-[`benchmarks/choice_at_scale.json`](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/benchmarks/choice_at_scale.json)). Each of the 118 requests
+([`benchmarks/choice_at_scale.py`](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/benchmarks/choice_at_scale.py), every choice in
+[`benchmarks/choice_at_scale.json`](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/benchmarks/choice_at_scale.json)). Each of the 118 requests
 was handed over in a set of 30, of 40 and of 50 candidates, filled with the operations the ranker
 finds closest, so the extra ones are the confusable ones; twelve choosers (Claude Opus 5.5) so
 that none saw a request twice; and each request compared with itself.
@@ -599,7 +613,7 @@ Three measurements say this is the right shape, and the third is the one that se
 
 | | |
 |---|---|
-| our ranking puts the answer in the top three | **61%** |
+| our ranking puts the answer in the top three | **60%** |
 | a model handed the same candidates and asked to *choose* gets its first pick right | **69%** |
 | the two labellers who wrote the ground truth agree **with each other** | **70%** |
 
@@ -619,7 +633,7 @@ adopted wholesale rather than judged one at a time. Two different strengths of e
 the file says which is which.
 
 **The two model figures are dated: the labels were written on 2026-08-28, against a catalogue of
-51 operations.** It now has 77, so for any request whose right answer is one of the 26 added since,
+51 operations.** It now has 78, so for any request whose right answer is one of the 27 added since,
 neither labeller *could* have been right — the answer was not in the catalogue to name. Measured on
 the first four requests a person has answered by hand, two of the four have both labellers wrong,
 and both of those two name operations that did not exist on the 28th. So "both labellers wrong" and
@@ -630,7 +644,7 @@ against a human answer where there is one.
 
 All three are over the same 118 requests, which matters: agreement measured over all 155 requests
 in the file is 68%, and the difference is the 21 pairs where both labellers agreed a request was
-unanswerable — true, and the easy half. Quoting that 68% beside a 61% computed over the 118 would
+unanswerable — true, and the easy half. Quoting that 68% beside a 60% computed over the 118 would
 be comparing two populations, which this table did for half a day.
 
 **The last row is a ceiling, not a baseline**, and the second row sits at it rather than below it.
@@ -687,9 +701,9 @@ catalog grows:
 
 | catalog size | BM25 found@3 | embeddings found@3 |
 |---|---|---|
-| 10 | 75% | 80% |
-| 30 | 68% | 57% |
-| 77 | 55% | 40% |
+| 10 | 73% | 82% |
+| 30 | 65% | 57% |
+| 78 | 55% | 40% |
 
 **This table used to say the opposite, and the reversal is the finding.** Published at 10/30/51
 it read 78/83, 47/65, 40/55 — embeddings ahead at every size — and the sentence under it said
@@ -740,11 +754,11 @@ number of features as the input. Those are structural properties of the operatio
 checkable against the code rather than declared by hand, and they separate the pairs a bag of
 words cannot: `spatial_join` from `overlay_layers`, `flow_accumulation` from `extract_streams`.
 That work is not done, and until it is, the honest claim is the measured one: the guarantee above
-holds at seventy-seven operations, not at eight hundred.
+holds at seventy-eight operations, not at eight hundred.
 
 **How an entry has to be written is a published specification**, not a convention:
-[`docs/catalog-entry-spec.md`](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/docs/catalog-entry-spec.md), with a normative
-[JSON Schema](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/schema/operation.schema.json) that every entry validates against in CI. Each field
+[`docs/catalog-entry-spec.md`](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/docs/catalog-entry-spec.md), with a normative
+[JSON Schema](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/schema/operation.schema.json) that every entry validates against in CI. Each field
 is there because a measurement said so — including the two that measured to nothing and are
 documented as such, because a spec that only reports what worked is an advertisement.
 
@@ -919,7 +933,9 @@ route to offer.
 Every tool that writes a dataset also writes `<output>.provenance.json` beside it and
 verifies its own work — CRS agreement, geometry validity, raster dimensions, count and
 extent invariants — recording the results in the manifest *before* raising anything, so
-the audit trail survives the error.
+the audit trail survives the error. A run that fails before writing, where an earlier
+run's dataset still sits at the path, leaves that run's record in place and writes its own
+as `<output>.failed.provenance.json`; the error names it.
 
 Verification runs on the way in as well. Before an operation touches your data, MapSmith
 checks the failures that produce *plausible* junk: an input with no CRS is refused
@@ -966,7 +982,7 @@ refusal to the remaining raster operations is on the roadmap below.
 
 ## See results inside the chat
 
-![MapSmith's interactive map panel rendered inside Claude Desktop: OSM basemap, buffer and zone layers, and per-layer provenance cards with verification status](https://raw.githubusercontent.com/mapsmith-ai/MapSmith/v0.8.0/docs/images/map-panel.png)
+![MapSmith's interactive map panel rendered inside Claude Desktop: OSM basemap, buffer and zone layers, and per-layer provenance cards with verification status](https://raw.githubusercontent.com/mapsmith-ai/MapSmith/v0.9.0/docs/images/map-panel.png)
 
 `preview_map` renders your layers on an interactive map panel *inside* the chat — pan,
 zoom, toggle layers, and read each layer's provenance card (operation, engine, and one of
@@ -1061,7 +1077,7 @@ layer that closes it is the right one: `INSTALL` and `LOAD` in a statement are r
 outright, because an `INSTALL` is an HTTPS fetch of a native binary run in this process on
 SQL a model wrote. Until 0.4.0 only the *implicit* forms were off, and an audit installed
 DuckDB's `aws` extension and read this machine's real cloud credentials back through a tool
-result — the whole story is in [SECURITY.md](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/SECURITY.md), including why none of the four
+result — the whole story is in [SECURITY.md](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/SECURITY.md), including why none of the four
 existing layers saw it. Extensions already loaded keep working, `spatial` included; to
 acquire others, name them where the agent cannot reach:
 **`MAPSMITH_ALLOW_EXTENSIONS=postgres,azure`** in the environment of the process that starts
@@ -1078,7 +1094,7 @@ GDAL parse attacker-chosen bytes in-process. Set **`MAPSMITH_ALLOW_REMOTE=1`** t
 refuses them regardless, since containment and "fetch whatever URL the model names" cannot
 both be true. The test suite asserts every branch by counting requests at a loopback server
 (`tests/test_duckdb_sandbox.py`). The full threat model —
-and what is explicitly *not* covered — is in [SECURITY.md](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/SECURITY.md).
+and what is explicitly *not* covered — is in [SECURITY.md](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/SECURITY.md).
 
 Fine print, because it changes how you deploy this: the path jail assumes a single trusted
 writer of the workspace filesystem (paths are resolved at check time, so a symlink swap by
@@ -1127,13 +1143,13 @@ each question rather than its position.
 ## We measured whether this actually helps
 
 Claims about agent performance are cheap, so
-[**docs/benchmarks.md**](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/docs/benchmarks.md) reports an A/B on
+[**docs/benchmarks.md**](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/docs/benchmarks.md) reports an A/B on
 [GABench](https://github.com/GeoX-Lab/GABench) — 57 executable GIS tasks over a
 133-tool server, scored by its deterministic evaluator — where the *only*
 variable is whether the agent's typed plan is validated before the solver runs.
 GABench is by Bo Yu et al. ([arXiv:2604.13888](https://arxiv.org/abs/2604.13888));
 the counts here are of the checkout we ran and differ from the paper's, which
-[docs/benchmarks.md](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/docs/benchmarks.md) states side by side.
+[docs/benchmarks.md](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/docs/benchmarks.md) states side by side.
 
 The honest headline is a **null result**, on a frontier model and on a small
 one, and the interesting part is why:
@@ -1161,7 +1177,7 @@ accuracy (parity on tool selection, measurably worse on ordering). One of those
 arms also refuted a conclusion this page had published two arms earlier; the
 correction is kept in place rather than edited away.
 
-The harness is in [`benchmarks/gabench-ab/`](https://github.com/mapsmith-ai/MapSmith/tree/v0.8.0/benchmarks/gabench-ab/), including
+The harness is in [`benchmarks/gabench-ab/`](https://github.com/mapsmith-ai/MapSmith/tree/v0.9.0/benchmarks/gabench-ab/), including
 the `split_analysis.py` that took our own win apart and the
 `rep_analysis.py` that bars every delta against a measured noise floor.
 
@@ -1190,7 +1206,7 @@ at all but a refusal: a raster and the sidecar beside it declare different geore
 both readings are GDAL behaving as documented, and the right move is to stop and say so
 rather than pick one. That family arrived on 2026-08-31, the morning after the list of
 twenty-seven was closed, and it is the reason a manifest can now name its `environment`
-(see the [changelog](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/CHANGELOG.md)).
+(see the [changelog](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/CHANGELOG.md)).
 
 The twenty-ninth family arrived on 2026-09-02 and is the one that reads oddly until you
 see it: a survey plot whose easement ring is wound the same way as its outline. A
@@ -1245,7 +1261,7 @@ suite outside:
 
 ## Notebook gallery
 
-Four executable walkthroughs in [`examples/`](https://github.com/mapsmith-ai/MapSmith/tree/v0.8.0/examples/): verified buffer+clip with
+Four executable walkthroughs in [`examples/`](https://github.com/mapsmith-ai/MapSmith/tree/v0.9.0/examples/): verified buffer+clip with
 provenance manifests, terrain and hydrology on a real 520×520 USGS DEM of **Mount St.
 Helens**, a deliberately wrong plan rejected before execution and then repaired, and
 **Copernicus** data on Monte Baldo — a DEM in degrees refused for slope and then reprojected,
@@ -1284,7 +1300,7 @@ detects it, converts the input first, and discloses the workaround in the manife
 - **You need an authenticated remote server today.** The Streamable HTTP transport has no
   authentication in this release: anyone who can reach the endpoint can run every tool
   against everything the process can see. Loopback or a trusted network only
-  ([SECURITY.md](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/SECURITY.md)).
+  ([SECURITY.md](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/SECURITY.md)).
 - **You want a sandbox for arbitrary agent code.** MapSmith confines paths and the SQL
   engine; there is no code-execution tool yet, and a path jail is not a container.
 - **You need cartography.** No styling, no layouts, no print composer. Outputs are
@@ -1297,14 +1313,14 @@ detects it, converts the input first, and discloses the workaround in the manife
   workspace — which is what the container runs with by default. DuckDB's own HTTP and S3
   filesystems stay off in every mode, so `read_parquet('s3://…')` does not work even with
   the opt-in: fetch the data down first, or run unconfined with remote reads on.
-- **You want the full breadth of a desktop GIS.** 77 operations, and a catalogue that tells
+- **You want the full breadth of a desktop GIS.** 78 operations, and a catalogue that tells
   the agent what does *not* exist yet rather than letting it improvise. The ~900 QGIS
   Processing algorithms are on the roadmap, not in the box.
 - **You expect plan validation to make a weak model strong.** Our own A/B says advisory
   validation upstream of an improvising solver does approximately nothing at aggregate
   level — and the enforced configuration MapSmith ships, measured afterwards, did not beat
   it on accuracy either. What enforcement buys is reproducibility
-  ([the numbers](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/docs/benchmarks.md)).
+  ([the numbers](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/docs/benchmarks.md)).
 - **You want us to debug your local geospatial toolchain.** Docker, or `uvx` where the
   wheels work, are the only supported paths; a hand-built native GDAL stack is not, on
   purpose.
@@ -1326,7 +1342,7 @@ Next, in the order we intend to do it. The linked items carry a written spec —
   under Apache-2.0. What it measures in MapSmith, including the seven defects it has sent back,
   has [a section of its own](#what-the-suite-found-in-mapsmith). #25 is closed against Argleton
   rather than left open here.
-- [ ] [Agent-loop repair](https://github.com/mapsmith-ai/MapSmith/issues/26): hand verification failures back to the agent as structured, actionable errors, with a bounded retry budget recorded in the manifest. Our [own measurements](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/docs/benchmarks.md) say the runtime error message is the information channel that works
+- [ ] [Agent-loop repair](https://github.com/mapsmith-ai/MapSmith/issues/26): hand verification failures back to the agent as structured, actionable errors, with a bounded retry budget recorded in the manifest. Our [own measurements](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/docs/benchmarks.md) say the runtime error message is the information channel that works
 - [ ] [Tool contracts that carry their own rules](https://github.com/mapsmith-ai/MapSmith/issues/27): argument constraints enforced *and* stated, and errors that name the rule rather than only the violation. The one intervention in our benchmark work that moved a metric past its noise floor
 - [ ] **A project brief for the requests that are a chain, not a call.** A third of real requests
   are not one operation: on the fifty hardest requests in our own set — the ones two model
@@ -1343,7 +1359,7 @@ Next, in the order we intend to do it. The linked items carry a written spec —
   validated plan — and every sentence in it has to trace back to a declared field, so a claim that
   traces back to nothing is a claim somebody invented, which is checkable by machine. And
   **MapSmith installs nothing**: the brief names what is missing and prints the command, and a
-  person runs it. The reason is dated — see [SECURITY.md](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/SECURITY.md) on why `INSTALL` in a SQL
+  person runs it. The reason is dated — see [SECURITY.md](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/SECURITY.md) on why `INSTALL` in a SQL
   statement is refused in both modes since 0.4.0. A server that installs on authorisation is the
   same shape with a consent dialog in front, and the consent comes from somebody who has just read
   persuasive prose written by the process asking for it.
@@ -1351,7 +1367,7 @@ Next, in the order we intend to do it. The linked items carry a written spec —
   Depends on the item above: the brief reads the rules the tools declare, so there is less to
   derive from without them.
 - [ ] [Satellite embeddings as a first-class input](https://github.com/mapsmith-ai/MapSmith/issues/24): per-zone embedding vectors (multiband zonal statistics) and similarity rasters against a reference location, over the open [AlphaEarth annual dataset](https://developers.google.com/earth-engine/guides/aef_on_gcs_readme) (CC-BY 4.0 COGs). Deterministic arithmetic on a raster — no model inference in MapSmith — with the tile, year and reference vector recorded in the manifest
-- [ ] Authenticated remote mode (OAuth on the existing Streamable HTTP transport) and [long-job progress via MCP Tasks](https://github.com/mapsmith-ai/MapSmith/issues/8). This is the item that closes the one limitation [SECURITY.md](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/SECURITY.md) declares outright: the HTTP transport has no authentication today
+- [ ] Authenticated remote mode (OAuth on the existing Streamable HTTP transport) and [long-job progress via MCP Tasks](https://github.com/mapsmith-ai/MapSmith/issues/8). This is the item that closes the one limitation [SECURITY.md](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/SECURITY.md) declares outright: the HTTP transport has no authentication today
 - [x] Slope and aspect (Whitebox, closed-form tested; geographic-CRS DEMs refused)
 - [x] Stream network extraction (Whitebox, from a flow-accumulation grid; the threshold and its unit recorded in the manifest)
 - [x] More terrain & hydrology: curvature (six kinds, the kind required because profile and plan answer opposite questions), flow direction (d8/rho8/dinf/fd8, with the direction-code **table** written into the manifest — the engine's own manual documents its default table backwards, so a name would not have been enough), Euclidean distance and IDW interpolation
@@ -1362,24 +1378,24 @@ Next, in the order we intend to do it. The linked items carry a written spec —
 
 ## License and project
 
-- MapSmith server and engines: **AGPL-3.0-or-later** (see [LICENSE](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/LICENSE))
+- MapSmith server and engines: **AGPL-3.0-or-later** (see [LICENSE](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/LICENSE))
 - Client SDK and tool-schema definitions (future `sdk/`): **Apache-2.0**
 
 You can self-host MapSmith freely, forever. If you modify it and offer it as a service, the
 AGPL asks you to share your changes — or [talk to us](mailto:mapsmith@proton.me) about a
 commercial license.
 
-Nothing here has been funded so far. [`funding.json`](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/funding.json) lists, in the
+Nothing here has been funded so far. [`funding.json`](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/funding.json) lists, in the
 [FLOSS/fund](https://fundingjson.org/) format, the two pieces of work money was asked for:
 a public suite of geospatial traps with hand-computable answers, and the provenance
 manifest as a specification other tools can implement. Both were built anyway, unfunded,
 and are archived with DOIs; the entries stay in the file marked inactive, so the estimate
 can be read against what came of it.
 
-Release notes are in [CHANGELOG.md](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/CHANGELOG.md), how to contribute in
-[CONTRIBUTING.md](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/CONTRIBUTING.md), how to report a vulnerability in
-[SECURITY.md](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/SECURITY.md). "MapSmith" is a trademark of the MapSmith project — see
-[TRADEMARKS.md](https://github.com/mapsmith-ai/MapSmith/blob/v0.8.0/TRADEMARKS.md). Updates:
+Release notes are in [CHANGELOG.md](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/CHANGELOG.md), how to contribute in
+[CONTRIBUTING.md](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/CONTRIBUTING.md), how to report a vulnerability in
+[SECURITY.md](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/SECURITY.md). "MapSmith" is a trademark of the MapSmith project — see
+[TRADEMARKS.md](https://github.com/mapsmith-ai/MapSmith/blob/v0.9.0/TRADEMARKS.md). Updates:
 [LinkedIn](https://www.linkedin.com/company/mapsmith) ·
 [X](https://x.com/mapsmith_ai) ·
 [Bluesky](https://bsky.app/profile/mapsmith.bsky.social).