microecs 0.3.2__tar.gz → 0.3.3__tar.gz

{microecs-0.3.2/microecs.egg-info → microecs-0.3.3}/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.4
 Name: microecs
-Version: 0.3.2
+Version: 0.3.3
 Summary: MicroECS: Minimal Entity Component System (ECS) in python and numpy
 Home-page: https://gitlab.com/meehai/microecs
 License: MIT
@@ -9,6 +9,14 @@ Description-Content-Type: text/markdown
 License-File: LICENSE.TXT
 Requires-Dist: numpy>=2.2.0
 Requires-Dist: loggez>=0.8
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: home-page
+Dynamic: license
+Dynamic: license-file
+Dynamic: requires-dist
+Dynamic: requires-python
+Dynamic: summary
 
 # MicroECS
 
@@ -16,24 +24,35 @@ Minimal (~300 LoC) Entity Component System in python and numpy. Examples also us
 
 Usage:
 
-- `pip install -r requirements.txt`
-- Sandbox: `python main.py`
-- Tests: `pytest test/`
+- Via pip: `pip install microecs`
+- From source code:
+```bash
+git clone https://gitlab.com/meehai/microecs      # clone the source code
+cd microecs                                       # go in the cloned directory
+python -m venv .venv && source .venv/bin/activate # make a virtual env, optional but useful
+python -m pip install -e .                        # install micro ecs in this virtual env
+python -m pytest test/                            # run the unit & integration tests to verify installation
+python examples/01-hello-world.py                 # run the basic hello world example (others in that dir)
+```
+
+Docs: [meehai.gitlab.io/microecs](https://meehai.gitlab.io/microecs/)
+
+## Relevant primitives: `Component`, `Pool`, `QueryResult`, `World`
 
-There are only 4 primitives (bottom up): `Component`, `Pool`, `QueryResult`, `World`:
+There are only 4 primitives (bottom up):
 
 - `Component` is a simple python dataclass holding only data. All entries must be numpy arrays with metadata fields: shape and dtype. We support 5 dtypes only: `int32`, `float32`, `bool`, `str` and `object`. A component with no fields is a valid **tag** for querying (e.g. `class Frozen(Component): pass`).
 - `Pool` is a simple 'archetype' dynamic array, holding entities of the same type (same set of components). Usses `Components` metadata to construct contiguous arrays for all entities of the same type.
-- `QueryResult` is a list of pools that match some query on all the entities of the `World`. It acts as a contiguous numpy-like container that implements numpy's `__array_function__` and `__array_ufunc__`. For all intents and purposes it should feel like a `(N, ...)` view over all the entities. If you need a numpy array (not all ops are supported, for e.g. indexing on the first axis), use `QueryResult.numpy()` (see [the field numpy contract](#the-field-_field-numpy-contract) below). It also exposes `entity_ids`: a flat `(N,)` array of the matched entities' ids, in the same pool-by-pool order as the fields, so you can `zip(qr.entity_ids, qr.position)` or feed an id back to `world.get_entity` / `world.remove_entity`.
+- `QueryResult` is a list of pools that match some query on all the entities of the `World`. It acts as a contiguous numpy-like container that implements numpy's interface. For all intents and purposes it should feel like a `(N, ...)` view over all selected entities. To get a proper numpy array, use `qr.numpy()`. To iterate over each entity in a query result (e.g. rendering), use `for eid, position in zip(qr.entity_ids, qr.position): ...`.
 - `World` is a manager of `Pools` and has an overview of all the entities in the scene. It also manages the migration of entities from one pool to the other. A `World` can also require extra metadata keys on every field via `World(extra_metadata=["serializable"])`, to enforce component-level behavior such as field serialization.
 
-Few relevant concepts:
+### Few relevant concepts:
 
 - `Pool` operates on array indices, while `World` operates on entity IDs (also integers). This allows seamless movement between pools while the high-level systems still working as intended.
 - All mutable operations on `World` are lazy. These are: `add_entity`, `remove_entity`, `add_component`, `remove_component`. They are added to a command buffer which is only executed when calling `world.update()`.
 - `Systems` are a convention, they are not part of this library. They can be defined at application level and act as hooks or callbacks. The `World` object doesn't need to know more than entities and components.
 
-Super simplified main loop structure:
+## Super simplified main loop structure
 
 ```python
 from typing import Callable
@@ -89,40 +108,22 @@ def main():
         rl.EndDrawing()
 ```
 
-## The field (`_Field`) numpy contract
-
-`qr.position` returns a `_Field`: a view over the matching pools that behaves like one `(N, *e)`
-numpy array (e.g. `(N, 2)` for a `(2,)` field). It applies each op **per pool** and stitches the
-result back. So the contract is exactly:
-
-> Any op that treats rows independently behaves **identically to numpy** on the concatenated
-> `(N, *e)` array — same values, same shape, and the same error on bad shapes.
+## Per-entity systems (e.g. rendering, foreign APIs)
 
-That covers elementwise math and ufuncs, broadcasting (every operand shape numpy accepts, and it
-raises on the ones numpy rejects), comparisons, `np.where` / `np.clip`, batched `np.matmul`,
-**feature-axis** reductions/indexing (`np.linalg.norm(qr.velocity, axis=1)`, `qr.pose[:, 0]`), and
-write-back broadcasting (`qr.position[:] = <scalar | (*e,) | (N, *e)>`). Pinned in
-`test/unit/test_field_numpy_parity.py`.
+Not every system vectorizes. A renderer calls a draw function per primitive; same for any per-entity foreign API. That's a per-entity system inside an otherwise-vectorized app. The cheapest way to iterate through all the entities is by using `zip()` on the `QueryResult` object on the fields you need, e.g:
 
-Edge cases worth knowing:
+```python
+qr = world.query(HasPosition, HasColor, HasRadius)
+for pos, color, radius in zip(qr.position, qr.color, qr.radius):     # one entity per step, fields aligned
+    rl.DrawCircle(int(pos[0]), int(pos[1]), float(radius[0]), color) # per-entity by necessity (no "draw all")
+```
 
-- **Not a full ndarray — these *raise*, never lie.** Entity-axis indexing beyond a single
-  `qr.f[i]` (`qr.f[:]`, `qr.f[2:4]`, `qr.f[mask]`, fancy), partial entity writes, and ndarray
-  methods/attrs (`.sum()`, `.mean()`, `.dtype`, `.ndim`, `.T`). Need any of these? Materialize
-  first with `qr.f.numpy()`.
-- **Axis-0 ops are per-pool, not global (footgun).** `np.sort` / `np.cumsum` / `np.sum` over
-  `axis=0` run *within each pool* and reset at pool boundaries — they do **not** see all entities
-  at once, so they differ from numpy. They're allowed, but if you want a global result, do
-  `qr.f.numpy()` first. A reduction that collapses the entity axis is rejected when its length no
-  longer matches the pool's row count.
-- **Operands must come from the same query.** Alignment is per-pool, not by flat index, so don't
-  mix a `_Field` from one `world.query(...)` into an op on another.
+There are other ways to extract a single entity e.g. `qr.some_field[i]` or `e = world.get_entity(eid)`, but these are slower in the general case and should be avoided in expensive loops, like rendering, even if they are a bit more ergonomic. See benchmark below.
 
-## Benchmark: ECS vs OOP
+<details>
+<summary> Microbenchmark: ECS vs OOP on a simple physics step </summary>
 
-The same physics step -- `pos += vel*dt` over N=100k entities split across 2 pools -- computed 8
-ways, all verified to produce the identical result. Reproduce with
-`python test/manual/perf/bench_ecs_vs_oop.py` (it prints `{mode: avg_seconds_per_step}`).
+Benchmark: we run the same physics step `pos += vel*dt` over N=100k entities split across 2 pools in various ways (ECS or OOP). All methods are verified to produce the identical result. Reproduce with `python test/manual/perf/bench_ecs_vs_oop.py` (it prints `{mode: avg_seconds_per_step}`).
 
 | pattern | ns/entity | vs OOP-scalar |
 |---|---:|---:|
@@ -135,8 +136,6 @@ ways, all verified to produce the identical result. Reproduce with
 | `micro-ecs-get-entity` — `world.get_entity(eid)` per entity | 1674 | 35× slower |
 | `micro-ecs-index` — `qr.f[i]` per entity (an `np.searchsorted` each) | 4573 | 95× slower |
 
-*Intel Core Ultra 7 165H, single-threaded, python 3.12 / numpy 2.2. ns/entity = seconds/step ÷ N.*
-
 Three things to take from it:
 
 1. **Vectorized wins big.** Batched ops (`_Field` or per-pool) run at 1–2 ns/entity — **27–52×
@@ -152,3 +151,26 @@ Three things to take from it:
 **Rule of thumb:** keep systems vectorized and push branches into `np.where` / `np.clip`. If a
 workload is *irreducibly* per-entity (data-dependent control flow), plain python objects beat
 microecs ~15× — use them there. microecs is the right tool for **vectorizable** simulation.
+
+</details>
+
+## How much are `Pool` and `QueryResult` numpy-like and corner cases
+
+Given `qr=world.query(A, B)`, then `qr.position` returns a `_Field`: a view over the matching pools that behaves like one contiguous-like `(N, *e)`
+numpy array (e.g. `(N, 2)` for a `(2,)` field). It applies each op **per pool** and stitches result back.
+
+That covers elementwise math and ufuncs (e.g. `np.where`, `np.linalg.norm(..., axis=1)` etc.), broadcasting (every operand shape numpy accepts, and it raises on the ones numpy rejects). See `test/unit/test_field_numpy_parity.py` for a whole set of operations comparing both.
+
+Edge cases worth knowing:
+
+- **Not a full ndarray — these raise, never lie.** Entity-axis indexing beyond a single
+  `qr.f[i]` (`qr.f[:]`, `qr.f[2:4]`, `qr.f[mask]`, fancy), partial entity writes, and ndarray
+  methods/attrs (`.sum()`, `.mean()`, `.dtype`, `.ndim`, `.T`). Need any of these? Materialize
+  first with `qr.f.numpy()`.
+- **Axis-0 ops are per-pool, not global (footgun).** `np.sort` / `np.cumsum` / `np.sum` over
+  `axis=0` run within each pool and reset at pool boundaries — they do **not** see all entities
+  at once, so they differ from numpy. They're allowed, but if you want a global result, do
+  `qr.f.numpy()` first. A reduction that collapses the entity axis is rejected when its length no
+  longer matches the pool's row count.
+- **Operands must come from the same query.** Alignment is per-pool, not by flat index, so don't
+  mix a `_Field` from one `world.query(...)` into an op on another.

microecs-0.3.2/PKG-INFO → microecs-0.3.3/README.md

@@ -1,39 +1,38 @@
-Metadata-Version: 2.1
-Name: microecs
-Version: 0.3.2
-Summary: MicroECS: Minimal Entity Component System (ECS) in python and numpy
-Home-page: https://gitlab.com/meehai/microecs
-License: MIT
-Requires-Python: >=3.12
-Description-Content-Type: text/markdown
-License-File: LICENSE.TXT
-Requires-Dist: numpy>=2.2.0
-Requires-Dist: loggez>=0.8
-
 # MicroECS
 
 Minimal (~300 LoC) Entity Component System in python and numpy. Examples also use raylib for rendering.
 
 Usage:
 
-- `pip install -r requirements.txt`
-- Sandbox: `python main.py`
-- Tests: `pytest test/`
+- Via pip: `pip install microecs`
+- From source code:
+```bash
+git clone https://gitlab.com/meehai/microecs      # clone the source code
+cd microecs                                       # go in the cloned directory
+python -m venv .venv && source .venv/bin/activate # make a virtual env, optional but useful
+python -m pip install -e .                        # install micro ecs in this virtual env
+python -m pytest test/                            # run the unit & integration tests to verify installation
+python examples/01-hello-world.py                 # run the basic hello world example (others in that dir)
+```
+
+Docs: [meehai.gitlab.io/microecs](https://meehai.gitlab.io/microecs/)
 
-There are only 4 primitives (bottom up): `Component`, `Pool`, `QueryResult`, `World`:
+## Relevant primitives: `Component`, `Pool`, `QueryResult`, `World`
+
+There are only 4 primitives (bottom up):
 
 - `Component` is a simple python dataclass holding only data. All entries must be numpy arrays with metadata fields: shape and dtype. We support 5 dtypes only: `int32`, `float32`, `bool`, `str` and `object`. A component with no fields is a valid **tag** for querying (e.g. `class Frozen(Component): pass`).
 - `Pool` is a simple 'archetype' dynamic array, holding entities of the same type (same set of components). Usses `Components` metadata to construct contiguous arrays for all entities of the same type.
-- `QueryResult` is a list of pools that match some query on all the entities of the `World`. It acts as a contiguous numpy-like container that implements numpy's `__array_function__` and `__array_ufunc__`. For all intents and purposes it should feel like a `(N, ...)` view over all the entities. If you need a numpy array (not all ops are supported, for e.g. indexing on the first axis), use `QueryResult.numpy()` (see [the field numpy contract](#the-field-_field-numpy-contract) below). It also exposes `entity_ids`: a flat `(N,)` array of the matched entities' ids, in the same pool-by-pool order as the fields, so you can `zip(qr.entity_ids, qr.position)` or feed an id back to `world.get_entity` / `world.remove_entity`.
+- `QueryResult` is a list of pools that match some query on all the entities of the `World`. It acts as a contiguous numpy-like container that implements numpy's interface. For all intents and purposes it should feel like a `(N, ...)` view over all selected entities. To get a proper numpy array, use `qr.numpy()`. To iterate over each entity in a query result (e.g. rendering), use `for eid, position in zip(qr.entity_ids, qr.position): ...`.
 - `World` is a manager of `Pools` and has an overview of all the entities in the scene. It also manages the migration of entities from one pool to the other. A `World` can also require extra metadata keys on every field via `World(extra_metadata=["serializable"])`, to enforce component-level behavior such as field serialization.
 
-Few relevant concepts:
+### Few relevant concepts:
 
 - `Pool` operates on array indices, while `World` operates on entity IDs (also integers). This allows seamless movement between pools while the high-level systems still working as intended.
 - All mutable operations on `World` are lazy. These are: `add_entity`, `remove_entity`, `add_component`, `remove_component`. They are added to a command buffer which is only executed when calling `world.update()`.
 - `Systems` are a convention, they are not part of this library. They can be defined at application level and act as hooks or callbacks. The `World` object doesn't need to know more than entities and components.
 
-Super simplified main loop structure:
+## Super simplified main loop structure
 
 ```python
 from typing import Callable
@@ -89,40 +88,22 @@ def main():
         rl.EndDrawing()
 ```
 
-## The field (`_Field`) numpy contract
-
-`qr.position` returns a `_Field`: a view over the matching pools that behaves like one `(N, *e)`
-numpy array (e.g. `(N, 2)` for a `(2,)` field). It applies each op **per pool** and stitches the
-result back. So the contract is exactly:
+## Per-entity systems (e.g. rendering, foreign APIs)
 
-> Any op that treats rows independently behaves **identically to numpy** on the concatenated
-> `(N, *e)` array — same values, same shape, and the same error on bad shapes.
+Not every system vectorizes. A renderer calls a draw function per primitive; same for any per-entity foreign API. That's a per-entity system inside an otherwise-vectorized app. The cheapest way to iterate through all the entities is by using `zip()` on the `QueryResult` object on the fields you need, e.g:
 
-That covers elementwise math and ufuncs, broadcasting (every operand shape numpy accepts, and it
-raises on the ones numpy rejects), comparisons, `np.where` / `np.clip`, batched `np.matmul`,
-**feature-axis** reductions/indexing (`np.linalg.norm(qr.velocity, axis=1)`, `qr.pose[:, 0]`), and
-write-back broadcasting (`qr.position[:] = <scalar | (*e,) | (N, *e)>`). Pinned in
-`test/unit/test_field_numpy_parity.py`.
-
-Edge cases worth knowing:
+```python
+qr = world.query(HasPosition, HasColor, HasRadius)
+for pos, color, radius in zip(qr.position, qr.color, qr.radius):     # one entity per step, fields aligned
+    rl.DrawCircle(int(pos[0]), int(pos[1]), float(radius[0]), color) # per-entity by necessity (no "draw all")
+```
 
-- **Not a full ndarray — these *raise*, never lie.** Entity-axis indexing beyond a single
-  `qr.f[i]` (`qr.f[:]`, `qr.f[2:4]`, `qr.f[mask]`, fancy), partial entity writes, and ndarray
-  methods/attrs (`.sum()`, `.mean()`, `.dtype`, `.ndim`, `.T`). Need any of these? Materialize
-  first with `qr.f.numpy()`.
-- **Axis-0 ops are per-pool, not global (footgun).** `np.sort` / `np.cumsum` / `np.sum` over
-  `axis=0` run *within each pool* and reset at pool boundaries — they do **not** see all entities
-  at once, so they differ from numpy. They're allowed, but if you want a global result, do
-  `qr.f.numpy()` first. A reduction that collapses the entity axis is rejected when its length no
-  longer matches the pool's row count.
-- **Operands must come from the same query.** Alignment is per-pool, not by flat index, so don't
-  mix a `_Field` from one `world.query(...)` into an op on another.
+There are other ways to extract a single entity e.g. `qr.some_field[i]` or `e = world.get_entity(eid)`, but these are slower in the general case and should be avoided in expensive loops, like rendering, even if they are a bit more ergonomic. See benchmark below.
 
-## Benchmark: ECS vs OOP
+<details>
+<summary> Microbenchmark: ECS vs OOP on a simple physics step </summary>
 
-The same physics step -- `pos += vel*dt` over N=100k entities split across 2 pools -- computed 8
-ways, all verified to produce the identical result. Reproduce with
-`python test/manual/perf/bench_ecs_vs_oop.py` (it prints `{mode: avg_seconds_per_step}`).
+Benchmark: we run the same physics step `pos += vel*dt` over N=100k entities split across 2 pools in various ways (ECS or OOP). All methods are verified to produce the identical result. Reproduce with `python test/manual/perf/bench_ecs_vs_oop.py` (it prints `{mode: avg_seconds_per_step}`).
 
 | pattern | ns/entity | vs OOP-scalar |
 |---|---:|---:|
@@ -135,8 +116,6 @@ ways, all verified to produce the identical result. Reproduce with
 | `micro-ecs-get-entity` — `world.get_entity(eid)` per entity | 1674 | 35× slower |
 | `micro-ecs-index` — `qr.f[i]` per entity (an `np.searchsorted` each) | 4573 | 95× slower |
 
-*Intel Core Ultra 7 165H, single-threaded, python 3.12 / numpy 2.2. ns/entity = seconds/step ÷ N.*
-
 Three things to take from it:
 
 1. **Vectorized wins big.** Batched ops (`_Field` or per-pool) run at 1–2 ns/entity — **27–52×
@@ -152,3 +131,26 @@ Three things to take from it:
 **Rule of thumb:** keep systems vectorized and push branches into `np.where` / `np.clip`. If a
 workload is *irreducibly* per-entity (data-dependent control flow), plain python objects beat
 microecs ~15× — use them there. microecs is the right tool for **vectorizable** simulation.
+
+</details>
+
+## How much are `Pool` and `QueryResult` numpy-like and corner cases
+
+Given `qr=world.query(A, B)`, then `qr.position` returns a `_Field`: a view over the matching pools that behaves like one contiguous-like `(N, *e)`
+numpy array (e.g. `(N, 2)` for a `(2,)` field). It applies each op **per pool** and stitches result back.
+
+That covers elementwise math and ufuncs (e.g. `np.where`, `np.linalg.norm(..., axis=1)` etc.), broadcasting (every operand shape numpy accepts, and it raises on the ones numpy rejects). See `test/unit/test_field_numpy_parity.py` for a whole set of operations comparing both.
+
+Edge cases worth knowing:
+
+- **Not a full ndarray — these raise, never lie.** Entity-axis indexing beyond a single
+  `qr.f[i]` (`qr.f[:]`, `qr.f[2:4]`, `qr.f[mask]`, fancy), partial entity writes, and ndarray
+  methods/attrs (`.sum()`, `.mean()`, `.dtype`, `.ndim`, `.T`). Need any of these? Materialize
+  first with `qr.f.numpy()`.
+- **Axis-0 ops are per-pool, not global (footgun).** `np.sort` / `np.cumsum` / `np.sum` over
+  `axis=0` run within each pool and reset at pool boundaries — they do **not** see all entities
+  at once, so they differ from numpy. They're allowed, but if you want a global result, do
+  `qr.f.numpy()` first. A reduction that collapses the entity axis is rejected when its length no
+  longer matches the pool's row count.
+- **Operands must come from the same query.** Alignment is per-pool, not by flat index, so don't
+  mix a `_Field` from one `world.query(...)` into an op on another.

{microecs-0.3.2 → microecs-0.3.3}/microecs/pool.py

@@ -17,7 +17,7 @@ class Pool:
         self.shapes = shapes
         self.dtypes = dtypes
 
-        self.data: dict[str, np.ndarray] = {} # the actual data is stored in a dynamic arrray, one per field
+        self.data: dict[str, np.ndarray] = {} # the actual data is stored in dict of dynamic arrrays, one per field
         self.size = 0
         self.capacity = Pool.INITIAL_CAPACITY
         for _field, shape, dtype in zip(fields, shapes, dtypes):

{microecs-0.3.2 → microecs-0.3.3}/microecs/query_result.py

@@ -91,6 +91,9 @@ class _Field(np.lib.mixins.NDArrayOperatorsMixin):
     def __len__(self):
         return self.len
 
+    def __repr__(self):
+        return f"[Field] Shape: {self.shape} (across {len(self.parts)} pools)"
+
 class QueryResult:
     """A list of pools seen as a contiguous view. Fields (qr.position) implement array interface to look like numpy"""
     def __init__(self, pool_list: list[Pool], field_shapes: dict[str, Shape], field_dtypes: dict[str, np.dtype],

{microecs-0.3.2 → microecs-0.3.3}/microecs/world.py

@@ -77,6 +77,12 @@ class World:
         components = self.pool_to_components[pool]
         return entity, components
 
+    def set_entity_data(self, entity_id: EntityId, field_name: str, value: np.ndarray):
+        """Sets the value of one specific entity's field given an enttiy id. Non-vectorized operation, use rarely"""
+        assert entity_id in self._eid_to_pool_ix, f"Entity id {entity_id} not found. Perhaps you didn't world.update()"
+        pool, pool_ix = self._eid_to_pool_ix[entity_id]
+        pool.data[field_name][pool_ix] = value
+
     def add_component(self, entity_id: EntityId, component: ComponentType, **kwargs):
         """Adds a component to an entity given its id. Component data is sent to kwargs. Lazy; call update()"""
         assert entity_id in self.live_ids, f"Entity id: {entity_id} not in {self.live_ids=}"

microecs-0.3.2/README.md → microecs-0.3.3/microecs.egg-info/PKG-INFO

@@ -1,27 +1,58 @@
+Metadata-Version: 2.4
+Name: microecs
+Version: 0.3.3
+Summary: MicroECS: Minimal Entity Component System (ECS) in python and numpy
+Home-page: https://gitlab.com/meehai/microecs
+License: MIT
+Requires-Python: >=3.12
+Description-Content-Type: text/markdown
+License-File: LICENSE.TXT
+Requires-Dist: numpy>=2.2.0
+Requires-Dist: loggez>=0.8
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: home-page
+Dynamic: license
+Dynamic: license-file
+Dynamic: requires-dist
+Dynamic: requires-python
+Dynamic: summary
+
 # MicroECS
 
 Minimal (~300 LoC) Entity Component System in python and numpy. Examples also use raylib for rendering.
 
 Usage:
 
-- `pip install -r requirements.txt`
-- Sandbox: `python main.py`
-- Tests: `pytest test/`
+- Via pip: `pip install microecs`
+- From source code:
+```bash
+git clone https://gitlab.com/meehai/microecs      # clone the source code
+cd microecs                                       # go in the cloned directory
+python -m venv .venv && source .venv/bin/activate # make a virtual env, optional but useful
+python -m pip install -e .                        # install micro ecs in this virtual env
+python -m pytest test/                            # run the unit & integration tests to verify installation
+python examples/01-hello-world.py                 # run the basic hello world example (others in that dir)
+```
+
+Docs: [meehai.gitlab.io/microecs](https://meehai.gitlab.io/microecs/)
 
-There are only 4 primitives (bottom up): `Component`, `Pool`, `QueryResult`, `World`:
+## Relevant primitives: `Component`, `Pool`, `QueryResult`, `World`
+
+There are only 4 primitives (bottom up):
 
 - `Component` is a simple python dataclass holding only data. All entries must be numpy arrays with metadata fields: shape and dtype. We support 5 dtypes only: `int32`, `float32`, `bool`, `str` and `object`. A component with no fields is a valid **tag** for querying (e.g. `class Frozen(Component): pass`).
 - `Pool` is a simple 'archetype' dynamic array, holding entities of the same type (same set of components). Usses `Components` metadata to construct contiguous arrays for all entities of the same type.
-- `QueryResult` is a list of pools that match some query on all the entities of the `World`. It acts as a contiguous numpy-like container that implements numpy's `__array_function__` and `__array_ufunc__`. For all intents and purposes it should feel like a `(N, ...)` view over all the entities. If you need a numpy array (not all ops are supported, for e.g. indexing on the first axis), use `QueryResult.numpy()` (see [the field numpy contract](#the-field-_field-numpy-contract) below). It also exposes `entity_ids`: a flat `(N,)` array of the matched entities' ids, in the same pool-by-pool order as the fields, so you can `zip(qr.entity_ids, qr.position)` or feed an id back to `world.get_entity` / `world.remove_entity`.
+- `QueryResult` is a list of pools that match some query on all the entities of the `World`. It acts as a contiguous numpy-like container that implements numpy's interface. For all intents and purposes it should feel like a `(N, ...)` view over all selected entities. To get a proper numpy array, use `qr.numpy()`. To iterate over each entity in a query result (e.g. rendering), use `for eid, position in zip(qr.entity_ids, qr.position): ...`.
 - `World` is a manager of `Pools` and has an overview of all the entities in the scene. It also manages the migration of entities from one pool to the other. A `World` can also require extra metadata keys on every field via `World(extra_metadata=["serializable"])`, to enforce component-level behavior such as field serialization.
 
-Few relevant concepts:
+### Few relevant concepts:
 
 - `Pool` operates on array indices, while `World` operates on entity IDs (also integers). This allows seamless movement between pools while the high-level systems still working as intended.
 - All mutable operations on `World` are lazy. These are: `add_entity`, `remove_entity`, `add_component`, `remove_component`. They are added to a command buffer which is only executed when calling `world.update()`.
 - `Systems` are a convention, they are not part of this library. They can be defined at application level and act as hooks or callbacks. The `World` object doesn't need to know more than entities and components.
 
-Super simplified main loop structure:
+## Super simplified main loop structure
 
 ```python
 from typing import Callable
@@ -77,40 +108,22 @@ def main():
         rl.EndDrawing()
 ```
 
-## The field (`_Field`) numpy contract
-
-`qr.position` returns a `_Field`: a view over the matching pools that behaves like one `(N, *e)`
-numpy array (e.g. `(N, 2)` for a `(2,)` field). It applies each op **per pool** and stitches the
-result back. So the contract is exactly:
+## Per-entity systems (e.g. rendering, foreign APIs)
 
-> Any op that treats rows independently behaves **identically to numpy** on the concatenated
-> `(N, *e)` array — same values, same shape, and the same error on bad shapes.
+Not every system vectorizes. A renderer calls a draw function per primitive; same for any per-entity foreign API. That's a per-entity system inside an otherwise-vectorized app. The cheapest way to iterate through all the entities is by using `zip()` on the `QueryResult` object on the fields you need, e.g:
 
-That covers elementwise math and ufuncs, broadcasting (every operand shape numpy accepts, and it
-raises on the ones numpy rejects), comparisons, `np.where` / `np.clip`, batched `np.matmul`,
-**feature-axis** reductions/indexing (`np.linalg.norm(qr.velocity, axis=1)`, `qr.pose[:, 0]`), and
-write-back broadcasting (`qr.position[:] = <scalar | (*e,) | (N, *e)>`). Pinned in
-`test/unit/test_field_numpy_parity.py`.
-
-Edge cases worth knowing:
+```python
+qr = world.query(HasPosition, HasColor, HasRadius)
+for pos, color, radius in zip(qr.position, qr.color, qr.radius):     # one entity per step, fields aligned
+    rl.DrawCircle(int(pos[0]), int(pos[1]), float(radius[0]), color) # per-entity by necessity (no "draw all")
+```
 
-- **Not a full ndarray — these *raise*, never lie.** Entity-axis indexing beyond a single
-  `qr.f[i]` (`qr.f[:]`, `qr.f[2:4]`, `qr.f[mask]`, fancy), partial entity writes, and ndarray
-  methods/attrs (`.sum()`, `.mean()`, `.dtype`, `.ndim`, `.T`). Need any of these? Materialize
-  first with `qr.f.numpy()`.
-- **Axis-0 ops are per-pool, not global (footgun).** `np.sort` / `np.cumsum` / `np.sum` over
-  `axis=0` run *within each pool* and reset at pool boundaries — they do **not** see all entities
-  at once, so they differ from numpy. They're allowed, but if you want a global result, do
-  `qr.f.numpy()` first. A reduction that collapses the entity axis is rejected when its length no
-  longer matches the pool's row count.
-- **Operands must come from the same query.** Alignment is per-pool, not by flat index, so don't
-  mix a `_Field` from one `world.query(...)` into an op on another.
+There are other ways to extract a single entity e.g. `qr.some_field[i]` or `e = world.get_entity(eid)`, but these are slower in the general case and should be avoided in expensive loops, like rendering, even if they are a bit more ergonomic. See benchmark below.
 
-## Benchmark: ECS vs OOP
+<details>
+<summary> Microbenchmark: ECS vs OOP on a simple physics step </summary>
 
-The same physics step -- `pos += vel*dt` over N=100k entities split across 2 pools -- computed 8
-ways, all verified to produce the identical result. Reproduce with
-`python test/manual/perf/bench_ecs_vs_oop.py` (it prints `{mode: avg_seconds_per_step}`).
+Benchmark: we run the same physics step `pos += vel*dt` over N=100k entities split across 2 pools in various ways (ECS or OOP). All methods are verified to produce the identical result. Reproduce with `python test/manual/perf/bench_ecs_vs_oop.py` (it prints `{mode: avg_seconds_per_step}`).
 
 | pattern | ns/entity | vs OOP-scalar |
 |---|---:|---:|
@@ -123,8 +136,6 @@ ways, all verified to produce the identical result. Reproduce with
 | `micro-ecs-get-entity` — `world.get_entity(eid)` per entity | 1674 | 35× slower |
 | `micro-ecs-index` — `qr.f[i]` per entity (an `np.searchsorted` each) | 4573 | 95× slower |
 
-*Intel Core Ultra 7 165H, single-threaded, python 3.12 / numpy 2.2. ns/entity = seconds/step ÷ N.*
-
 Three things to take from it:
 
 1. **Vectorized wins big.** Batched ops (`_Field` or per-pool) run at 1–2 ns/entity — **27–52×
@@ -140,3 +151,26 @@ Three things to take from it:
 **Rule of thumb:** keep systems vectorized and push branches into `np.where` / `np.clip`. If a
 workload is *irreducibly* per-entity (data-dependent control flow), plain python objects beat
 microecs ~15× — use them there. microecs is the right tool for **vectorizable** simulation.
+
+</details>
+
+## How much are `Pool` and `QueryResult` numpy-like and corner cases
+
+Given `qr=world.query(A, B)`, then `qr.position` returns a `_Field`: a view over the matching pools that behaves like one contiguous-like `(N, *e)`
+numpy array (e.g. `(N, 2)` for a `(2,)` field). It applies each op **per pool** and stitches result back.
+
+That covers elementwise math and ufuncs (e.g. `np.where`, `np.linalg.norm(..., axis=1)` etc.), broadcasting (every operand shape numpy accepts, and it raises on the ones numpy rejects). See `test/unit/test_field_numpy_parity.py` for a whole set of operations comparing both.
+
+Edge cases worth knowing:
+
+- **Not a full ndarray — these raise, never lie.** Entity-axis indexing beyond a single
+  `qr.f[i]` (`qr.f[:]`, `qr.f[2:4]`, `qr.f[mask]`, fancy), partial entity writes, and ndarray
+  methods/attrs (`.sum()`, `.mean()`, `.dtype`, `.ndim`, `.T`). Need any of these? Materialize
+  first with `qr.f.numpy()`.
+- **Axis-0 ops are per-pool, not global (footgun).** `np.sort` / `np.cumsum` / `np.sum` over
+  `axis=0` run within each pool and reset at pool boundaries — they do **not** see all entities
+  at once, so they differ from numpy. They're allowed, but if you want a global result, do
+  `qr.f.numpy()` first. A reduction that collapses the entity axis is rejected when its length no
+  longer matches the pool's row count.
+- **Operands must come from the same query.** Alignment is per-pool, not by flat index, so don't
+  mix a `_Field` from one `world.query(...)` into an op on another.

{microecs-0.3.2 → microecs-0.3.3}/setup.py

@@ -3,7 +3,7 @@ from pathlib import Path
 from setuptools import setup, find_packages
 
 NAME = "microecs"
-VERSION = "0.3.2"
+VERSION = "0.3.3"
 DESCRIPTION = "MicroECS: Minimal Entity Component System (ECS) in python and numpy"
 URL = "https://gitlab.com/meehai/microecs"