{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Execute basic operations on the tree\n", "\n", "Hyperiax is designed for fast and easy-to-use tree-level executions. In this notebook, you will know the general workflow, including defining your operations, executing them using Hyperiax's engine, visualize the results, and a simple way to integrate all the parts using Hyperiax's prebuilt modules.\n", "\n", "For better illustration, we shall use an concrete example in phylogenetic analysis: estimating inner nodes of a phylogenetic tree, given:\n", "- observations of the leaf nodes\n", "- knowledge of the tree topology, including number of nodes, the edges connecting them and the edge lengths\n", "\n", "This is a basic problem in the field of _phylogenetic comparative methods_. We may be particularly interested in the root node estimation, which can be interpreted as the phylogenetic mean of the observations, where we shall use a very simple estimator the inner nodes: **mean estimator**, that is, any inner node will be estimated as the mean of its children weighted by its edge length:\n", "$$\n", "\\hat{x} = \\sum^k_{i=1} \\frac{1}{w_i}\\cdot c_i\n", "$$\n", "where $\\hat{x}$ is the estimation of node $x$, $w_i$ is the edge length of the $i$-th child node, and $c_i$ is the child's value. We set all the node values are 2-dimensional vectors, which allows for nicer visualizations, but you can of course use Hyperiax to deal with other high-dimensional data.\n", "\n", "In practice, we create trees from real data. However, in this notebook, we are not going to study a real-life example. Instead, we will create a hypothetical tree, and of course, we can manually assign values to each node. But we will show a faster method: simulating the tree from the root node, based on the Brownian motion model. Specifically, an edge of length $T$ corresponds to a Brownian motion running for a duration of $T$. \n", "\n", "**Content of the notebook**\n", "- Create the tree with a chosen topology and simulate its node values, based on the Brownian motion model\n", "- Estimate inner nodes based on the simulated tree\n", "- Compute the estimations more conveniently using `prebuilts` modules" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "import hyperiax\n", "from hyperiax.execution import OrderedExecutor\n", "from hyperiax.models import UpLambdaReducer, DownLambda, UpLambda\n", "from hyperiax.models.functional import pass_up\n", "from hyperiax.tree.topology import symmetric_topology\n", "from hyperiax.tree import HypTree\n", "from hyperiax.plotting import plot_tree_text, plot_tree_2d_scatter\n", "\n", "import jax\n", "from jax import numpy as jnp\n", "from jax.random import PRNGKey, split\n", "from matplotlib import pyplot as plt" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Simulate the tree (downwards)\n", "\n", "Recall briefly last notebook, where we learn that a `topology` is needed for Hyperiax to create the tree, this `topology` contains no actual data, and only serves as a representation of the data structure we intend to work on.\n", "\n", "We first create a symmetric tree with `height=4` and `degree=3`, which means the tree will have `4+1=5` layers, and each node has `3` children." ] }, { "cell_type": "markdown", "metadata": {}, "source": [] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " None\n", " ┌────────────────────────────────────────────┼────────────────────────────────────────────┐\n", " None None None \n", " ┌──────────────┼──────────────┐ ┌──────────────┼──────────────┐ ┌──────────────┼──────────────┐ \n", " None None None None None None None None None \n", " ┌────┼────┐ ┌────┼────┐ ┌────┼────┐ ┌────┼────┐ ┌────┼────┐ ┌────┼────┐ ┌────┼────┐ ┌────┼────┐ ┌────┼────┐ \n", "None None None None None None None None None None None None None None None None None None None None None None None None None None None\n" ] } ], "source": [ "topology = symmetric_topology(height=3, degree=3)\n", "plot_tree_text(topology)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The topology meets our expectation. Now we need to specify which kind of data we intend to store in the tree. Hyperiax allows for fast computation using any backend supported by _JAX_, and if the operations are suitable, _JAX_ will automatically remove all the computation to the GPUs (if available) and runs them at an extraordinary speed. Therefore, all the data store in the `HypTree` is stored in the format of `jax.numpy.ndarray`, which is the _Numpy_-like array with _JAX_ backend. To assign the data, we need to let Hyperiax know the shape of coming data first, which is for initializing buffer. But don't worry, all you need to do is just to pass the shape as tuples, and Hyperiax will take over the rest.\n", "\n", "In our case, we need to assign three properties to each node:\n", "- `edge_length`: as we use in the last notebook, we will use it to determine the scale of Brownian motions along the edges, with shape of `(1,)`\n", "- `value`: the data value stored in each node, in this case, is a 2-d vector with shape of `(2, )`\n", "- `noise`: the noise (for example, from the observations) in each node, with the same shape as `value`" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "tree = HypTree(topology)\n", "tree.add_property('noise', shape=(2,))\n", "tree.add_property('edge_length', shape=(1,))\n", "tree.add_property('value', shape=(2,))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "When calling `add_property()` method of `HypTree`, Hyperiax will initialize the a contiguous buffer in the memory with given shapes and number of nodes, with the BFS order in `tree.data`, and all the actual values are initialized to `0`.\n", "\n", "We first set the edge lengths to be exponentially decaying with the depth, like we did in the previous notebook" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [], "source": [ "tree.data['edge_length'] = jnp.exp(-jnp.log(2) * tree.node_depths)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We might also need to deal with some randomness, for example, the `noise` of each node. With the help of contiguous buffer created by Hyperiax, we can directly sample values for the entire tree, instead of do it one node at a time. This should be the preferred method, except the underlying distribution is node-dependent.\n", "\n", "Now, we can do the same for the 2-dimensional values that we intend to store in `noise`:" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [], "source": [ "key = PRNGKey(0)\n", "key, subkey = split(key)\n", "tree.data['noise'] = jax.random.normal(subkey, shape=tree.data['noise'].shape)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "After we initialize the tree, we need to define a function that runs downwards to simulate the Brownian motions along the edges. We shall call it `down` function for short in the result of the notebook. Generally, a `down` function should have the following features:\n", "- Defined on a edge $(u, v)$, where $u$ is the source node and $v$ is the target node.\n", "- Used to compute the new value of $v$ based on the current states (can only depends on $u$ or $v$ or both)\n", "\n", "With this in mind, go back to our Brownian case. The goal of our Brownian simulation is to simulate all the node values except the root, and the value of every node is sample from a Brownian motion starts from its parent node and runs for the time of the distance from the root to the evaluated node. If we denote the edge length as $T$, the value of evaluated root $v$ should follow the distribution of $\\mathcal{N}(v; u, T\\cdot\\mathbf{I})$. In practice, $v = u + \\sqrt{T}\\cdot\\epsilon$, where $\\epsilon\\sim\\mathcal{N}(0,\\mathbf{I})$ and can be generated using `jax.random.normal()` for example, or we can just use our previously defined `noise`.\n", "\n", "Before we start to define the `down` function, we need to know a bit about how Hyperiax treats the `down` function: In Hyperiax, functions on a tree always consider the _current_ node. When referring to properties such as `noise`, it means the `noise` stored in the _current_ node. To define custom operations on a tree, we can query the necessary data within nodes by specifying parameter names. In this case, we need both the `noise` and the `edge_length` in the _current_ node. Additionally, we need the result of calculations from the parent node. In hyperiax, values stored in the _parent_ node are prefixed with `parent_`. For this computation, we need `parent_value` (the result of the computation in the parent node).\n", "\n", "Now we are ready to define the `down` function. The implementation of the `down` function involves the following steps:\n", "\n", "1. **Extract the necessary parameters**:\n", " - `noise`: the noise value in the _current_ node.\n", " - `edge_length`: the edge length stored in the _current_ node (this is by convention the length of the edge connecting the _current_ and _parent_ node)\n", " - `parent_value`: the result of the computation in the _parent_ node.\n", "\n", "2. **Compute the new value for the child node**:\n", " - Calculate the new value as `sqrt(edge_length) * noise + parent_value`.\n", "\n", "3. **Return the new value**:\n", " - Return a dictionary specifying the `value` parameter to overwrite in the child node $v$. Values that are not being changed, such as `edge_length`, can be omitted.\n", "\n", "Hyperiax will greedily distribute your workloads, using vectorization among other techniques. This means that any `down` function is expected to work on _batched_ data. In our example, the input shapes will be as follows;\n", "\n", "- `noise`: `(b, 2)`.\n", "- `edge_length`: `(b, 1)`.\n", "\n", "Similarly, we need to return a _batched_ output, in this case `value`, should have the shape `(b, 2)`, where `b` is some batch size determined by hyperiax.\n", "\n", "If the concept of batching seems uncomfortable to you, we highly recommend using [`jax.vmap`](https://jax.readthedocs.io/en/latest/_autosummary/jax.vmap.html) which can do the batching for you automatically." ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [], "source": [ "@jax.jit\n", "def down(noise, edge_length,parent_value, **args):\n", " return {'value': jnp.sqrt(edge_length) * noise + parent_value}" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Here, we set the root value to be the `noise` we sampled. All the data is stored in BFS order in the buffers, this means that the root is always the first element." ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [], "source": [ "tree.data['value'] = tree.data['value'].at[0].set(tree.data['noise'][0])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Now we create both an **executor** and a **model**. The **model** object contains all user code related to the operations we wish to run on a tree. The **executor** object schedules execution within the tree, and is in charge of setting the data in the correct format, as well as batching things together and making sure everything runs in the correct order.\n", "\n", "You are welcome to define your operations as an entire class, in cases where a constructor is needed. In this example we use the anonymous **lambda models** in `hyperiax.models.lambdamodels.py`, which enables you to pass tree operations as arguments, such that they are ready to be executed on a tree directly." ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [], "source": [ "downmodel = DownLambda(down_fn=down)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We have now created a `downmodel`, specified by the `down` function that will generate Brownian motion samples on a tree. The only thing left to do is give this **model** to an executor, which will run the code in the correct order on the tree. In this case we use the `OrderedExecutor`, since order is important for our operation. This executor will work on models going in either the up or down direction on a tree." ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [], "source": [ "exe = OrderedExecutor(downmodel)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "With these objects created, we can now run our `down` function on the tree! Be careful as these operations are **INPLACE**. Because of this, we reccomend [idempotent](https://en.wikipedia.org/wiki/Idempotence) operations like here, where the result is stored in `value`, instead of overwriting `noise`(notice how this would give a different result for each execution)." ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [], "source": [ "exe.down(tree)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Besides topology visualizations, Hyperiax also provides avenues of displaying the underlying data in the tree. When dealing with $2$ dimensional vectors, we can perform a classic scatter plot, but with the tree topology overlaid.\n", "\n", "Note that passing in a pyplot axis object is optional, if you do not wish to use your own, then it can be omitted.\n", "\n", "Using these two plots, we can see how the `down` function infact seems to correlate our nodes!" ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig, ax = plt.subplots(1,2, figsize = (10,4))\n", "for a in ax:\n", " a.set_xlim(-2, 4)\n", " a.set_ylim(-2, 2)\n", "plot_tree_2d_scatter(tree, 'value', ax=ax[0])\n", "ax[0].set_title('Value')\n", "plot_tree_2d_scatter(tree, 'noise', ax=ax[1])\n", "ax[1].set_title('Noise');" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The green node represents the root, and the light blue nodes represent the leafs." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Estimate the inner nodes (upwards)\n", "\n", "So far, we assume we know the value of root, together with the model that runs from top to bottom. However, in most of the real cases, we know about the leaf nodes, and want to inference the root and inner nodes. In phylogenetics, the leaf nodes are the species we observe at current time, and the root and inner nodes represents their ancestors.\n", "\n", "Now we going to mask all the data on the inner nodes of our simulated tree, and only keeps the leaf nodes. To do this, `HypTree` has a boolean mask that indicates whether or not a node is a obtainable leaf via `tree.is_leaf`. Therefore, we can add another property `estimated_value` to the tree, with the same shape as `value`, and initialized as `0`. Then we index the leaf nodes by slicing through `is_leaf` and manually set them to be the simulated leaf `value`: " ] }, { "cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [], "source": [ "tree.add_property('estimated_value', shape=(2,))\n", "leaf_data = tree.data['value'][tree.is_leaf]\n", "tree.data['estimated_value'] = tree.data['estimated_value'].at[tree.is_leaf].set(leaf_data)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We have all the observations on the leaf nodes, we want to estimate the nodes above. The direction is upward, from bottom to top. Thus, we need a `up` function just like the `down` function we defined before. But unlike the downward pass, where the message is passing through a one-to-one pair $(u, v)$, the upward pass will aggregate all the information from child nodes to the parent node, so it is a many-to-one correspondence. Therefore beside `up` function, we need to define another two addons to facilitate Hyperiax to handle this: `reductions` parameter and `transform` function:\n", "* The `reductions` parameter of the `UpLambda` class, determines how each property should be reduced when multiple are incoming from the _child_ nodes. In this case we have `sum` for both arguments. This means that after a message is sent from each child toward its parent, each property is first summed before the `transform` operation in the _parent_.\n", "* The `transform` function processes the reduced incoming messages of the _child_ nodes, and treats the receiving node of the messages as the _current_ node. This means that we need to use the prefix `child_` when getting the reduced values of the children.\n", "* The `up` function then takes the value in a _current_ node, and forms a message that is passed toward its _parent_." ] }, { "cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [], "source": [ "def up(estimated_value, edge_length, **kwargs):\n", " return {'weighted_value': estimated_value/edge_length, \n", " 'inverse_edge_length':1/edge_length}\n", " \n", "def transform(child_weighted_value,child_inverse_edge_length, **kwargs):\n", " return {'estimated_value': child_weighted_value/child_inverse_edge_length}\n", "\n", "upmodel = UpLambdaReducer(up, \n", " transform, \n", " reductions={\n", " 'weighted_value': 'sum', \n", " 'inverse_edge_length': 'sum'\n", " })" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Since upward pass is more complicated than the downward pass, it is worthy to explain it more here. We first summarize what we expect about `up`, `reduction` and `transform`:\n", "- **Up Operation (`up` function)**:\n", " - **Purpose**: To generate a message at _each_ node that will be passed to its parent.\n", " - **Inputs**: Values specific to the current node.\n", " - **Outputs**: A message containing the processed values to be sent to the parent node.\n", "\n", "- **Reduction (`reductions` parameter)**:\n", " - **Purpose**: To aggregate messages from _multiple_ child nodes.\n", " - **Inputs**: Messages from the child nodes.\n", " - **Outputs**: A single aggregated message that combines the values from all child nodes using a specified reduction method (e.g., sum).\n", "\n", "- **Transform Operation (`transform` function)**:\n", " - **Purpose**: To process the aggregated message at the _parent_ node.\n", " - **Inputs**: Aggregated messages from the child nodes aswell as the values in the _current_ node.\n", " - **Outputs**: A transformed value for the parent node.\n", "\n", "Then we can explain the above code in details:\n", "\n", "**1. Up Operation (`up` function)**\n", "\n", "The `up` function takes the value at the current node and forms a message that will be passed to the parent node. \n", "\n", "When hyperiax executes functions on a tree, it always has a sense of the _current_ node. In hyperiax terminology, when referring to a specific property, such as `estimated_value`, it means the `estimated_value` stored in the _current_ node.\n", "\n", "- **Parameters**:\n", " - `estimated_value`: The value associated with the current node.\n", " - `edge_length`: The length of the edge connecting the current node to its parent.\n", " - `**kwargs`: Any additional arguments that might be needed.\n", "- **Returns**: A dictionary with two properties:\n", " - `weighted_value`: Calculated as `estimated_value / edge_length`.\n", " - `inverse_edge_length`: Calculated as `1 / edge_length`.\n", "\n", "In this case, with $w_i=1/l_i$ with $l_i$ being the edge length in node $i$. We can write the inner node estimation $p$ as a sum over children $c_i$\n", "\n", "$$p=\\frac{1}{\\sum w_i}\\sum c_iw_i$$\n", "\n", "The key is to observe that we can compute $\\sum w_i$ and $\\sum c_iw_i$ independently, where we use the sum as a reduction and $w_i$ and $w_ic_i$ are the messages being passed up\n", "\n", "**2. Transform Operation (`transform` function)**\n", "\n", "The `transform` function processes the aggregated message at the parent node. It uses the reduced values from the child nodes to compute a new value for the current (parent) node.\n", "\n", "- **Parameters**:\n", " - `child_weighted_value`: The aggregated `weighted_value` from all child nodes.\n", " - `child_inverse_edge_length`: The aggregated `inverse_edge_length` from all child nodes.\n", " - `**kwargs`: Any additional arguments that might be needed.\n", "- **Returns**: A dictionary with one property:\n", " - `estimated_value`: Calculated as `child_weighted_value / child_inverse_edge_length`.\n", "\n", "**3. Reduction (`reductions` parameter)**\n", "\n", "The reduction step determines how to combine multiple messages received from child nodes. In this example, we use the sum operation for both `weighted_value` and `inverse_edge_length` properties.\n", "\n", "- **`weighted_value`**: The messages' `weighted_value` properties from all child nodes are summed.\n", "- **`inverse_edge_length`**: The messages' `inverse_edge_length` properties from all child nodes are summed." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "After understanding what all these are used for, we can put them all together and warp them by the **lambda model** as we did for the downward pass. To set up the upward pass, we define an `UpLambda` model with the `up` and `transform` functions, and the specified reductions.\n", "\n", "- **`UpLambda`**: Initializes the upward pass model with the `up` and `transform` functions and the reduction rules.\n", "- **`OrderedExecutor`**: Executes the upward pass using the defined model.\n", "\n", "This setup ensures that values propagate correctly from the leaves to the root, are aggregated at each step, and transformed to provide the final values at each node. Then we can execute the computation and let Hyperiax take over all the rest." ] }, { "cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "13.3 μs ± 178 ns per loop (mean ± std. dev. of 7 runs, 100,000 loops each)\n" ] } ], "source": [ "upward_exe = OrderedExecutor(upmodel)\n", "upward_exe.up(tree)\n", "%timeit upward_exe.up(tree)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We can of course visualize the estimated results:" ] }, { "cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig, ax = plt.subplots(1, 2, figsize = (10,4))\n", "for a in ax:\n", " a.set_xlim(-2, 4)\n", " a.set_ylim(-2, 2)\n", "plot_tree_2d_scatter(tree, 'value', ax=ax[0])\n", "ax[0].set_title('Value')\n", "plot_tree_2d_scatter(tree, 'estimated_value', ax=ax[1])\n", "ax[1].set_title('Estimated Value');" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We can see that the estimated values follow the same distribution as the simulated ones, which is reasonable since we used the same model for downward and upward passes." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "**Getting around with `reductions`**\n", "\n", "We also allow for a simpler interface in the up-direction at a cost. If your tree has a constant or zero number of children for all nodes, and you are willing to do some upfront work `(precompute_child_gathers=True)`, then you can get around the reduce operation by simply specifying an `up` function that can access data in child nodes and the current node. The returned values are to be set in the current node.\n", "\n", "The way it handles children is a bit different. In our example, our `value`s in the children are 2 dimensional vectors, and since each node in our topology has exactly `3` or `0` children, we say that the base of the tree is `3`. We can now query the child values by giving `child_estimated_value` as a parameter to our up function. This will have the shape `[batch, base, dimension]` where in our case `batch` is determined by hyperiax, `base=3` from the tree, and `dimension=2` because we are working in 2d. We now need to aggregate these such that we return something with the shape `[batch, dimension]` as the estimated value. In this example, we do it by a weighted mean, using the edge lengths as weights." ] }, { "cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [], "source": [ "pre_tree = HypTree(topology, precompute_child_gathers=True)\n", "pre_tree.data = tree.data.copy()" ] }, { "cell_type": "code", "execution_count": 17, "metadata": {}, "outputs": [], "source": [ "# use precomputed child gathers to accelerate the upward pass\n", "def up(child_estimated_value, child_edge_length, **kwargs):\n", " inv_el = 1/child_edge_length\n", " normalizer = inv_el.sum(axis=1)\n", "\n", " return {'estimated_value': jnp.einsum('bcd,bc1->bd', child_estimated_value, inv_el)/normalizer}" ] }, { "cell_type": "code", "execution_count": 18, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "10.2 μs ± 298 ns per loop (mean ± std. dev. of 7 runs, 100,000 loops each)\n", "True\n" ] } ], "source": [ "upmodel = UpLambda(up_fn=up)\n", "upmodelexe = OrderedExecutor(upmodel)\n", "res = upmodelexe.up(pre_tree)\n", "%timeit upmodelexe.up(pre_tree)\n", "print((pre_tree.data['estimated_value'] == tree.data['estimated_value']).all())" ] }, { "cell_type": "code", "execution_count": 19, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig, ax = plt.subplots(1,2, figsize = (10,4))\n", "for a in ax:\n", " a.set_xlim(-2, 4)\n", " a.set_ylim(-2, 2)\n", "plot_tree_2d_scatter(tree, 'estimated_value', ax=ax[0])\n", "ax[0].set_title('Estimated Value')\n", "plot_tree_2d_scatter(pre_tree, 'estimated_value', ax=ax[1])\n", "ax[1].set_title('Estimated Value (precomputed)');" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We can see the precomputed tree gets exactly the same results as doing it in the normal way, but runs 25% faster (9.64μs compared to 12.9μs). However, such precomputation is only available when the parents have constant number of children, while in a more general case, the previous way of defining `up` is recommended." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Using `prebuilts` to finish them all in once\n", "\n", "Alternatively, we can use the cool `hyperiax.models.prebuilts` to do the same computations as above, but more conveniently!" ] }, { "cell_type": "code", "execution_count": 20, "metadata": {}, "outputs": [], "source": [ "from hyperiax.models.prebuilts import PhyloMeanModel" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The `PhyloMeanExecutor` expects leaf nodes to all contain a field called `estimated_value`. Since we already know the exact values here, we can simply set them directly" ] }, { "cell_type": "code", "execution_count": 21, "metadata": {}, "outputs": [], "source": [ "pme = PhyloMeanModel()\n", "prebuilt_exe = OrderedExecutor(pme)" ] }, { "cell_type": "code", "execution_count": 22, "metadata": {}, "outputs": [], "source": [ "prebuilt_exe.up(tree)" ] }, { "cell_type": "code", "execution_count": 23, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig, ax = plt.subplots(1,2, figsize = (10,4))\n", "for a in ax:\n", " a.set_xlim(-2, 4)\n", " a.set_ylim(-2, 2)\n", "plot_tree_2d_scatter(tree, 'value', ax=ax[0])\n", "ax[0].set_title('Value')\n", "plot_tree_2d_scatter(tree, 'noise', ax=ax[1])\n", "ax[1].set_title('Noise');" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "In this case, we don't need to care about defining `up`, `reductions` or `transform` anymore, just run `prebuilt_exe.up()`, which is super easy and cool!\n", "\n", "## Summary\n", "Now you know how to define a basic mean estimation operation in Hyperiax, and execute it downwards or upwards. In the next notebook, we will show a more complicated, but also very interesting task in biological morphometry: aligning two species' shapes on the phylogenetic tree, and how Hyperiax can handle this in an easy and efficient way." ] } ], "metadata": { "kernelspec": { "display_name": "hyperiax_dev", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.11.4" } }, "nbformat": 4, "nbformat_minor": 2 }