feat(events)!: advancedFilter becomes a filter expression

AGENTS.md

@@ -185,7 +185,7 @@ Names the refactors removed — `externalIdPrefix`, `metadataKey`/`metadataValue
 
 Related resources are one field, not two: both `Event` and `EventFilter` carry `relatedResources`, an array of the backend's `IdCollection` (`[{"id": "34"}, {"externalId": "sensor_abc"}]`, modelled by `IdAndExtId`). An entry may name a resource by id, external id, or both; the backend resolves the missing side and returns both. `dataSetId` on the filter uses the same shape. There are no aliases for the retired flat `relatedResourceIds` / `relatedResourceExternalIds` arrays.
 
-`EventFilterForm` + `AdvancedEventFilter` remain the richer style for events; some advanced-filter endpoints are not yet wired up server-side and are tested only via serde round-trips.
+`EventFilterForm` additionally carries `advancedFilter`, a **string** holding a boolean expression in the api's PostgreSQL-flavoured filter language — `type NOT LIKE 'pump' AND (subType = 'water' OR subType = 'gas')`. It replaced a nested and/or/not `Filter` tree in 0.4.0. Three of that tree's nine variants (`range`, `containsAny`, `containsAll`) never had a server-side binding and returned 500, which is part of why it went. The expression is parsed and validated by the api, not here, so an invalid one comes back as a 400 carrying an offset and often a corrected expression.
 
 #### Sorting and paging
 

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "intellistream-datahub-sdk"
-version = "0.3.0"
+version = "0.4.0"
 edition = "2021"
 rust-version = "1.85"
 description = "Async Rust SDK for the IntelliStream DataHub REST API: time series and datapoints, events, resources and relationship edges, files, datasets, labels and subscriptions."

datahub_python_bindings/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "datahub_python_bindings"
-version = "0.3.0"
+version = "0.4.0"
 edition = "2024"
 rust-version = "1.85"
 description = "PyO3 bindings for the IntelliStream DataHub SDK, published to PyPI as intellistream-datahub-sdk."

datahub_python_bindings/pyproject.toml

@@ -11,7 +11,7 @@ exclude = ["**/__pycache__/**"]
 [project]
 name = "intellistream-datahub-sdk"
 
-version = "0.3.0"
+version = "0.4.0"
 description = "Python bindings for the IntelliStream DataHub SDK: time series and datapoints, events, resources and relationship edges, files, datasets, labels and subscriptions."
 readme = "README.md"
 authors = [{ name = "IntelliStream AS", email = "support@intellistream.ai" }]

datahub_python_bindings/python/intellistream_datahub_sdk/__init__.pyi

@@ -899,6 +899,7 @@ class EventsServiceSync:
         sort_by: SortBy | None = None,
         sort_order: str | None = None,
         cursor: str | None = None,
+        advanced_filter: str | None = None,
     ) -> Page:
         """Pass either ``filter=`` or the individual criteria keywords; passing both is a
         ``TypeError``. Paging is always given here rather than on the filter, so one filter can be
@@ -954,6 +955,7 @@ class EventsServiceAsync:
         sort_by: SortBy | None = None,
         sort_order: str | None = None,
         cursor: str | None = None,
+        advanced_filter: str | None = None,
     ) -> Page:
         """Pass either ``filter=`` or the individual criteria keywords; passing both is a
         ``TypeError``. Paging is always given here rather than on the filter, so one filter can be

datahub_python_bindings/src/events/async_service.rs

@@ -124,7 +124,8 @@ impl PyEventsServiceAsync {
     #[pyo3(signature = (filter=None, external_id=None, source=None, r#type=None, sub_type=None,
                         status=None, data_set_id=None, event_time=None, metadata=None,
                         related_resources=None, created_time=None, last_updated_time=None,
-                        limit=None, sort_by=None, sort_order=None, cursor=None))]
+                        limit=None, sort_by=None, sort_order=None, cursor=None,
+                        advanced_filter=None))]
     #[allow(clippy::too_many_arguments)]
     fn filter<'py>(
         &self,
@@ -145,11 +146,12 @@ impl PyEventsServiceAsync {
         sort_by: Option<crate::StringOrList>,
         sort_order: Option<String>,
         cursor: Option<String>,
+        advanced_filter: Option<String>,
     ) -> PyResult<Bound<'py, PyAny>> {
         let form = crate::events::event_filter_form(
             filter, external_id, source, r#type, sub_type, status, data_set_id, event_time,
             metadata, related_resources, created_time, last_updated_time, limit, sort_by,
-            sort_order, cursor,
+            sort_order, cursor, advanced_filter,
         )?;
         let service = self.api_service.clone();
 

datahub_python_bindings/src/events/mod.rs

@@ -86,6 +86,7 @@ pub fn event_filter_form(
     sort_by: Option<StringOrList>,
     sort_order: Option<String>,
     cursor: Option<String>,
+    advanced_filter: Option<String>,
 ) -> PyResult<EventFilterForm> {
     let any_keyword = external_id.is_some()
         || source.is_some()
@@ -131,6 +132,13 @@ pub fn event_filter_form(
     if let Some(cursor) = cursor {
         form.set_cursor(cursor);
     }
+    // A boolean expression in the api's filter language, e.g.
+    //   type NOT LIKE 'pump' AND (subType = 'water' OR subType = 'gas')
+    // Parsed and validated by the api, so an invalid one comes back as a 400 that carries an
+    // offset and usually a corrected expression.
+    if let Some(advanced_filter) = advanced_filter {
+        form.set_advanced_filter(advanced_filter);
+    }
     Ok(form.build())
 }
 

datahub_python_bindings/src/events/sync_service.rs

@@ -108,7 +108,8 @@ impl PyEventsServiceSync {
     #[pyo3(signature = (filter=None, external_id=None, source=None, r#type=None, sub_type=None,
                         status=None, data_set_id=None, event_time=None, metadata=None,
                         related_resources=None, created_time=None, last_updated_time=None,
-                        limit=None, sort_by=None, sort_order=None, cursor=None))]
+                        limit=None, sort_by=None, sort_order=None, cursor=None,
+                        advanced_filter=None))]
     #[allow(clippy::too_many_arguments)]
     fn filter<'py>(
         &self,
@@ -129,11 +130,12 @@ impl PyEventsServiceSync {
         sort_by: Option<crate::StringOrList>,
         sort_order: Option<String>,
         cursor: Option<String>,
+        advanced_filter: Option<String>,
     ) -> PyResult<crate::PyPage> {
         let form = crate::events::event_filter_form(
             filter, external_id, source, r#type, sub_type, status, data_set_id, event_time,
             metadata, related_resources, created_time, last_updated_time, limit, sort_by,
-            sort_order, cursor,
+            sort_order, cursor, advanced_filter,
         )?;
         let service = self.api_service.clone();
 

docs/api-coverage.md

@@ -90,8 +90,9 @@ vocabularies). All take a `limit` param, default 1000, capped at 10000.
 Covered: `byids`, `search`, `fetch-related`, `create`, `update`, `delete`.
 
 Missing: `GET /resources/{id}` (single fetch by numeric id), `POST /resources/filter` (structured
-filtering — note the SDK ships `AdvancedFilter` machinery in `src/filters.rs` with no resource
-endpoint wired to it), `POST /resources/fetch-nearest`.
+filtering), `POST /resources/fetch-nearest`. Note that events' `advancedFilter` is now a plain
+expression string and carries no reusable type, so a resource equivalent would start from the
+api's filter language rather than from anything in `src/filters.rs`.
 
 ### `/timeseries` — 11 of 12
 
@@ -149,7 +150,8 @@ knowing, found while testing against a live backend:
 3. ~~`GET /files/download/{id}`~~ — **done**, along with the rest of `/files`.
 4. **`/edges` service** — needed for any relationship management that isn't a resource-create side
    effect, and for discovering relationship types.
-5. **`/resources/filter`** — the `AdvancedFilter` types already exist; only the call is missing.
+5. **`/resources/filter`** — needs both the call and whatever filter shape the api settles on for
+   resources; the events filter language is a string, with no Rust-side types to reuse.
 6. **`/policies` and `/governance`** — newer surfaces; wire up when the platform needs them.
 7. Lower value: events vocabulary enumeration, `/stats`, `/tenant/features`,
    `recommend-value-type`, the live datapoint WebSocket.