Changelog

Changelog#

1.0.0 (pre-release)#

⚠️ Version 1.0.0 makes multiple changes which might brake your code when upgrading. Please read this how-to guide on how to upgrade to version 1.0.0.

🧩 New features#

  • A larger and more flexible set of synapse dynamics (#748, @michaeldeistler):

from jaxley.synapses import CurrentSynapse, ConductanceSynapse, DynamicSynapse, IonotropicSynapse

  • Fire channels now implement a surrogate gradient (#735, @alexpejovic)

  • New SpikeSynapse added, which detects Fire channel spikes (#735, @alexpejovic)

  • New AlphaSynapse added (#764, @michaeldeistler)

  • Channels can now have an init_params method. You can then initialize parameters depending on one another with:

cell.init_params()

  • Methods to simplify assigning recordings to their location in the cell or network (#750, @michaeldeistler):

recs = jx.integrate(cell)
cell.write_recordings(recs)
print(cell.branch(0).comp(1).recording("v"))  # Only the recording in branch 0, comp 1.

  • Synapses can now use the states and parameters of pre- and post-synaptic compartments (#765, @michaeldeistler)

API changes#

  • Synapses now take different arguments in order to be able to fetch the pre- and post-synaptic states of the network (#735, @alexpejovic)

  • Removed the TanhConductanceSynapse (#748, @michaeldeistler). Replace it with:

import jax.numpy as jnp
from jaxley.synapses import ConductanceSynapse
connect(..., ConductanceSynapse(jnp.tanh))

  • Removed the TanhRateSynapse (#748, @michaeldeistler). Replace it with:

import jax.numpy as jnp
from jaxley.synapses import CurrentSynapse
connect(..., CurrentSynapse(jnp.tanh))

  • Synapses have a new API (#765, @michaeldeistler). The update_states method and the compute_current method should both receive all of the following arguments, in that order:

self,
synapse_states: dict[str, Array],
synapse_params: dict[str, Array],
pre_voltage: Array,
post_voltage: Array,
pre_states: dict[str, Array],
post_states: dict[str, Array],
pre_params: dict[str, Array],
post_params: dict[str, Array],
delta_t: float,

  • Channels have a new API (#766, @michaeldeistler). The update_states method and the compute_current method should both receive all of the following arguments, in that order:

self,
channel_states: dict[str, Array],
channel_params: dict[str, Array],
voltage: Array,
delta_t: float,

  • The default Na, K, and Leak channels have been changed (#766, @michaeldeistler). To recover the old channels, do:

pip install jaxley-mech

from jaxley_mech.channels.pospischil import Na, K, Leak

  • All Pospischil type channels have been moved to the jaxley-mech repository (#766, @michaeldeistler). To get access to these channels, do:

pip install jaxley-mech

from jaxley_mech.channels.pospischil import Na, K, Leak, CaL, CaT, M

  • cell.recordings has been renamed to cell.rec_info (#750, @michaeldeistler)

  • Remove solve_inf_gate_exponential. Replace with exponential_euler (#766, @michaeldeistler)

📚 Documentation#

  • How-to guide on upgrading to version 1.0 (#766, @michaeldeistler)

  • New tutorial added for simple SNNs (#735, @alexpejovic)

  • Improved documentation for synapses (#748, @michaeldeistler)

  • How-to guide on connecting synapses to pre-defined spike trains (#764, @michaeldeistler)

  • Improved documentation for channels (#766, @michaeldeistler)

0.13.0#

🧩 New features#

  • Add utilities to process the all_states dictionary such that it only contains “true” ODE states (i.e., it removes branchpoint states, membrane states that are NaN because the channel does not exist there, and ionic currents). In addition, the utilities can flatten the states to a vector (#719, @matthijspals, @michaeldeistler):

import jaxley as jx
from jaxley.integrate import build_init_and_step_fn
from jaxley.utils.dynamics import build_dynamic_state_utils

cell = jx.Cell()
params = cell.get_parameters()

init_fn, step_fn = build_init_and_step_fn(cell)
remove_observables, add_observables, flatten, unflatten = build_dynamic_state_utils(cell)

all_states, all_params = init_fn(params)

dynamic_states = flatten(remove_observables(all_states))
recovered_all_states = add_observables(unflatten(dynamic_states), all_params, delta_t=0.025)

📚 Documentation#

  • Improved documentation for build_init_and_step_fn (#719, @matthijspals, @michaeldeistler)

  • Improved examples in documentation (#759, @michaeldeistler)

  • Manually remove some functions from the documentation (#759, @michaeldeistler)

0.12.0#

🧩 New features#

  • Exponential Euler solver (#743, @michaeldeistler):

jx.integrate(cell, solver="exp_euler")

This solver is slow on CPU, but can be very performant on GPU, especially when length, radius, axial resistivity, and capacitance do not change across simulations (i.e., they are no parameters). In that case, the transition matrix for exponential Euler can be precomputed (and it is no longer computed at every call to jx.integrate):

cell.customize_solver_exp_euler(
    exp_euler_transition=cell.build_exp_euler_transition_matrix(delta_t)
)  # Pre-compute the update matrix. Has to be rerun for new values of radius, l, ra, C.
jx.integrate(cell, solver="exp_euler")

  • Forward Euler solver for branched morphologies (#743, @michaeldeistler).

🛠️ Internal updates#

  • separate getting the currents from get_all_states() (#727, @michaeldeistler). To restore the previous behaviour, do:

states = cell.get_all_states(pstate)
states = cell.append_channel_currents_to_states(states, all_params, delta_t)

🐛 Bug fixes#

  • bugfix for cell.vis(..., type="morph") (#725, @michaeldeistler, thanks to Elisabeth Galyo for reporting)

📚 Documentation#

  • Added example usage to many user-facing Module functions (#716, @alexpejovic)

  • Update GPU installation instructions to use CUDA 13 (#732, @michaeldeistler)

  • Update citation (#739, @michaeldeistler)

  • new how-to guide on choosing a solver (#743, @michaeldeistler)

  • fixes for documentation rendering (#743, @michaeldeistler)

0.11.5#

🐛 Bug fixes#

  • bugfix for .delete() when multiple channels have the same current_name or a shared parameter/state (#713, @michaeldeistler)

  • safe softplus, use linear function above certain threshold. This avoids an unwanted clipping operation due to the save_exp (#714 @matthijspals)

📚 Documentation#

  • typo fixes for several tutorial notebooks (#721, @michaeldeistler, thanks @martricks for reporting)

  • update tutorial on building channel and synapse models: now includes pitfalls and recommendations (#723, @michaeldeistler)

0.11.4#

🐛 Bug fixes#

  • bugfix for indexing when init_states() is run on a jx.Network (#711, @michaeldeistler)

📚 Documentation#

  • add an example on fitting a morphologically detailed cell with gradient descent (#705, @michaeldeistler)

0.11.3#

🛠️ Internal updates#

  • follow jax typing practices with Array and ArrayLike (#693, @alexpejovic)

🐛 Bug fixes#

  • fix for networks that mix point neurons and morphologically detailed neurons (#702, @michaeldeistler)

  • carry over groups from constituents of a module (e.g., jx.Cell groups get carried over to jx.Network) (#703, @michaeldeistler)

0.11.2#

🐛 Bug fixes#

  • Bugfix for Networks on GPU: since v0.9.0, networks had been very slow on GPU because the voltage equations of cells had been processed in sequence, not in parallel. This is now solved, giving a large speed-up for networks consisting of many cells (#691, @michaeldeistler, thanks to @VENOM314 for reporting)

📚 Documentation#

  • Remove all content from the old mkdocs documentation website (#689, @michaeldeistler)

0.11.1#

🐛 Bug fixes#

  • bugfix for set_ncomp() when the cell consists of a single branch (#686, @michaeldeistler)

🛠️ Internal updates#

  • fix all typos in the codebase by using the typos project (#682, @alexpejovic)

0.11.0#

🧩 New features#

  • simple conductance synapse added (#659, @kyralianaka)

📚 Documentation#

  • add a how-to guide on converting NMODL files to Jaxley, see here (#669, @michaeldeistler, special thanks to @r-makarov for building the tool)

🛠️ Internal updates#

  • changes to how the membrane area from SWC files is computed when the radius within a compartment is not constant. This fix can have an impact on simulation results. The updated computation of membrane area matches that of the NEURON simulator (#662, @michaeldeistler, thanks to @VENOM314 for reporting).

  • remove the custom SWC reader (which had been deprecated in v0.9.0, #662, @michaeldeistler).

  • fix bug in the .ncomp attribute after set_ncomp() had been run (#676, @manuelgloeckler)

0.10.0#

🧩 New features#

  • functionality to compute the pathwise distance between compartments (#648, @michaeldeistler):

from jaxley.morphology import distance_pathwise
path_dists = distance_pathwise(cell.soma.branch(0).comp(0), cell)
cell.nodes["path_dist_from_soma"] = path_dists

🐛 Bug fixes#

  • fixed synapse recording indices to be within type (#643, @kyralianaka)

  • Fix inheriting from a Module #590 (#642, @jnsbck)

🛠️ Internal updates#

  • module.distance() is now deprecated in favor of jx.morphology.distance_direct() (#648, @michaeldeistler)

0.9.0#

✨ Highlights#

  • This PR implements a new solver, which is now used by default (#625, @michaeldeistler). The new solver has the following advantages:

    • Much lower runtime. Across several morphologies, we get a 20% runtime speedup on CPU and a 50% runtime speedup on GPU.

    • Almost zero compile time. We get a 50x compile time speedup on CPU and a 3x compile time speedup on GPU.

  • Utility to delete parts of a morphology (#612, @michaeldeistler):

from jaxley.morphology import morph_delete
cell = morph_delete(cell.branch([1, 2]))

  • Utility to connect two cells into a single cell (#612, @michaeldeistler):

from jaxley.morphology import morph_connect
cell = morph_connect(cell1.branch(1).loc(0.0), cell2.branch(2).loc(1.0))

🧩 New features#

  • the default SWC reader has changed. To use the previous SWC reader, run jx.read_swc(..., backend="custom"). However, note that we will remove this reader in the future. If the new SWC reader is causing issues for you, please open an issue (#612, @michaeldeistler)

  • radiuses are now integrated across SWC coordinates, not interpolated (#612, @michaeldeistler)

  • remove pin of JAX version. New JAX versions (JAX>=0.6.0) resolve slow CPU runtime, see here (#623, @michaeldeistler)

  • running the d_lambda rule is now much faster, see the how-to guide (#625, @michaeldeistler)

📚 Documentation#

  • Introduce the how-to guide on the website (#612, @michaeldeistler)

  • reorganize the advanced tutorials into subgroups (#612, @michaeldeistler)

  • split the morphology handling tutorials into two notebooks (#612, @michaeldeistler)

🛠️ Internal updates#

  • improvements to graph-backend for more flexibility in modifying morphologies (#613, @michaeldeistler)

  • jx.read_swc() now ignores type_id > 4 by default. This ensures compatibility with flywire, which highjacks type_id > 4 to indicate synaptic contacts (#612, @michaeldeistler)

  • remove root compartment for SWC files (#613, @michaeldeistler)

  • enable traversing compartmentalized graph for optimizing solve order (#613, @michaeldeistler)

  • ._comp_edges are being tracked in the View (#621, @michaeldeistler)

  • introduce ._branchpoints_ attribute and track in the View (#612, @michaeldeistler)

  • pre_global_comp_index and post_global_comp_index are no longer part of .edges by default. To get them, do net.copy_node_property_to_edges("global_comp_index") (#625, @michaeldeistler)

🐛 Bug fixes#

  • ChainTransform forward now working as mirror to inverse (#628, @kyralianaka)

  • allow data_set with vectors of values (#606, @chaseking)

🎉 New Contributors#

  • @chaseking made their first contribution in #606

0.8.2#

  • enable intersections of groups, e.g. net.exc.soma (#608, @michaeldeistler)

  • allow ion diffusion to be jitted (#609, @michaeldeistler)

0.8.1#

  • fixups to tutorial notebooks (#604, @michaeldeistler)

  • remove delete_channel() (#604, @michaeldeistler)

0.8.0#

🧩 New features#

  • add leaky integrate-and-fire neurons (#564, @jnsbck), Izhikevich neurons, and rate-based neurons (#601, @michaeldeistler)

🛠️ Minor updates#

  • make delta and v_th in IonotropicSynapse trainable parameters (#599, @jnsbck)

  • make random postsnaptic compartment selection optional in connectivity functions (#489, @kyralianaka)

🐛 Bug fixes#

  • Fix bug for groups when .set_ncomp was run (#587, @michaeldeistler)

  • allow .distance to be jitted (#603, @michaeldeistler)

0.7.0#

🧩 New Features#

  • Allow ion diffusion with cell.diffuse() and add tutorials (#438, @michaeldeistler):

from jaxley.channels import CaNernstReversal
from jaxley.pumps import CaFaradayConcentrationChange

cell.insert(CaFaradayConcentrationChange())
cell.insert(CaNernstReversal())
cell.diffuse("CaCon_i")
cell.set("axial_diffusion_CaCon_i", 1.0)

  • Introduce ion pumps (#438, @michaeldeistler)

🛠️ Minor changes#

  • rename delete_channel() to delete() (#438, @michaeldeistler)

🐛 Bug fixes#

  • Fix for simulation of morphologies with inhomogeneous numbers of compartments (#438, @michaeldeistler)

  • Bugfix for types assigned by SWC reader when soma is traced by a single point (#582, @Kartik-Sama, @michaeldeistler).

Code health#

  • add new release workflow (#588, @jnsbck)

  • update FAQ and tutorials (#593, @michaeldeistler)

  • make random post compartment selection optional in connectivity functions (#489, @kyralianaka)

🎉 New Contributors#

  • @Kartik-Sama made their first contribution in #582

0.6.2#

  • also remove nseg from tutorials (#580, @jnsbck)

0.6.1#

  • Fixup for failing readthedocs build of v0.6.0 (#579, @michaeldeistler)

  • Deprecate nseg in favor of ncomp.

0.6.0#

Pin of JAX version#

Installing Jaxley will no longer install the newest version of JAX. We realized that, on CPU, with version jax==0.4.32 or newer, simulation time in Jaxley is 10x slower and gradient time is 5x slower as compared to older versions of JAX. Newer versions of JAX can be made equally fast as older versions by setting os.environ['XLA_FLAGS'] = '--xla_cpu_use_thunk_runtime=false' at the beginning of your jupyter notebook (#570, @michaeldeistler).

🧩 New Features#

  • Add ability to record synaptic currents (#523, @ntolley). Recordings can be turned on with

net.record("i_IonotropicSynapse")

  • refactor plotting (#539, @jnsbck).

    • rm networkx dependency

    • add Network.arrange_in_layers

    • disentangle moving of cells and plotting in Network.vis. To get the same as net.vis(layers=[3,3]), one now has to do:

net.arrange_in_layers([3,3])
net.vis()

  • Allow parameter sharing for groups of different sizes, i.e. due to inhomogeneous numbers of compartments or for synapses with the same (pre-)synaptic parameters but different numbers of post-synaptic partners. (#514, @jnsbck)

  • Add jaxley.io.graph for exporting and importing of jaxley modules to and from networkx graph objects (#355, @jnsbck).

    • Adds a new (and improved) SWC reader, which is more flexible and should also be easier to extend in the future.

    from jaxley.io.graph import swc_to_graph, from_graph
graph = swc_to_graph(fname)
# do something to the swc graph, i.e. prune it
pruned_graph = do_something_to_graph(graph)
cell = from_graph(pruned_graph, ncomp=4)

    • Adds a new to_graph method for jaxley modules, which exports a module to a networkX graph. This allows to seamlessly work with networkX’s graph manipulation or visualization functions.

    • "graph" can now also be selected as a backend in the read_swc.

    • See the improved SWC reader tutorial for more details.

🛠️ Code Health#

  • changelog added to CI (#537, #558, @jnsbck)

  • Add regression tests and supporting workflows for maintaining baselines (#475, #546, @jnsbck).

    • Regression tests can be triggered by commenting “/test_regression” on a PR.

    • Regression tests can be done locally by running NEW_BASELINE=1 pytest -m regression i.e. on main and then pytest -m regression on feature, which will produce a test report (printed to the console and saved to .txt).

  • Allow inspecting the version via import jaxley as jx; print(jx.__version__) (#577, @michaeldeistler).

🐛 Bug fixes#

  • Fixed inconsistency with type assertions arising due to numpy functions returning different dtypes on platforms like Windows (#567, @Kartik-Sama)

🎉 New Contributors#

  • @ntolley made their first contribution in #523

0.5.0#

🛠️ API changes#

  • Synapse views no longer exist (#447, #453, @jnsbck). Previous code such as

net.IonotropicSynapse.set("IonotropicSynapse_s", 0.2)

must be updated to:

net.set("IonotropicSynapse_s", 0.2)

For a more detailed tutorial on how to index synapses, see this new tutorial.

  • Throughout the codebase, we renamed any occurrence of seg (for segment) to comp (for compartment). The most notable user-facing changes are:

    • branch = jx.Branch(comp, ncomp=4)

    • cell = jx.read_swc(fname, ncomp=4)

  • New defaults for the SWC reader with jx.read_swc(). By default, we now have assign_groups=True (previously False) and max_branch_len=None (previously 300.0).

  • We renamed .view to .nodes, e.g., cell.branch(0).nodes (#447, #453, @jnsbck).

  • We renamed _update_nodes_with_xyz() to compute_compartment_centers() (#520, @jnsbck)

  • We updated the way in which transformations are built (#455, @manuelgloeckler). Previous code such as

tf = jx.ParamTransform(
    lower={"radius": 0.1, "length": 2.0},
    lower={"radius": 3.0, "length": 20.0},
)

must be updated to:

from jaxley.optimize.transforms import ParamTransform, SigmoidTransform
transforms = [
    {"radius": SigmoidTransform(lower=0.1, upper=3.0)},
    {"length": SigmoidTransform(lower=2.0, upper=20.0)},
]
tf = jt.ParamTransform(transforms)

🧩 New features#

  • Added a new delete_channel() method (#521, @jnsbck)

  • Allow to write trainables to the module (#470, @michaeldeistler):

net.make_trainable("radius")
params = net.get_parameters()
net.write_trainables(params)

  • Expose the step function to allow for fine-grained simulation (#466, @manuelgloeckler)

  • More flexible and thorough viewing (#447, #453, @jnsbck)

  • Boolean indexing for cells, branches, and comps (@494, @jnsbck):

r_greater_1 = net.nodes.groupby("global_cell_index")["radius"].mean() > 1
net[r_greater_1].nodes.vis()

  • check if recordings are empty (#460, @deezer257)

  • enable clamp to be jitted and vmapped with data_clamp() (#374, @kyralianaka)

🐛 Bug fixes#

  • allow for cells that were read from swc to be pickled (#525, @jnsbck)

  • fix units of compute_current() in channels (#461, @michaeldeistler)

  • fix issues with plotting when the morphology has a different number of compartments (#513, @jnsbck)

📚 Documentation#

  • new tutorial on synapse indexing (#464, @michaeldeistler, @zinaStef)

  • new tutorial on parameter sharing (#464, @michaeldeistler, @zinaStef)

  • new tutorial on modules and views (#493, @jnsbck)

  • improved tutorial on building channel models (#473, @simoneeb)

  • get rid of tensorflow dependency by defining our simple dataloader in the tutorial (#484, @jnsbck)

  • new FAQ about rate-based networks (#531, @michaeldeistler)

Code health#

  • refactor tests with fixtures (@479, #499, @fabioseel, @jnsbck)

  • make several attributes private (#495, @ntolley)

  • move read_swc.py to new io folder (#524, @jnsbck)

  • faster testing for SWC and plotting (#479, @fabioseel)

  • automated tests to check if tutorials can be run (#480, @jnsbck)

  • add helpers to deprecate functions and kwargs (#516, @jnsbck)

🎉 New Contributors#

  • @simoneeb made their first contribution in #473

  • @zinaStef made their first contribution in #464

  • @fabioseel made their first contribution in #479

  • @deezer257 made their first contribution in #460

0.4.0#

🧩 New features#

  • Changing the number of compartments: cell.branch(0).set_ncomp(4) (#436, #440, #445, @michaeldeistler, @jnsbck)

  • New options for plotting: cell.vis(type='comp') and cell.vis(type='morph') (#432, #437, @jnsbck)

  • Speed optimization for jx.integrate(..., voltage_solver="jaxley.stone") (#442, @michaeldeistler)

📚 Documentation#

v0.3.0#

🧩 New features#

  • New solver: jx.integrate(..., voltage_solver="jax.sparse") which has very low compile time (#418, @michaeldeistler)

  • Support for different number of compartments per branch at initialization (modifying the number of compartments after initialization is not yet supported, #418, #426, @michaeldeistler)

🐛 Bug fixes#

  • Bugfix for capacitances and their interplay with axial conductances (Thanks @Tunenip, #426, @michaeldeistler)

  • Bugfixes for tutorials on website

v0.2.1#

  • Bugfix for using states in init_state of the channel (#421, @michaeldeistler)

  • Bugfix for tutorial on building custom channels (#421, @michaeldeistler)

  • Add jaxley-mech as dev dependency to use it in testing (#421, @michaeldeistler)

v0.2.0#

🧩 New features#

  • Cranck-Nicolson solver (#413, @michaeldeistler)

  • Forward Euler solver for compartments and branches (#413, @michaeldeistler)

  • Add option to access states in channel.init_state (#416, @michaeldeistler)

🐛 Bug fixes#

  • Bugfix for interpolation of x, y, z values (#411, @jnsbck)

v0.1.2#

  • Minor fixups for README, website, and PyPI landing page

v0.1.1#

  • Minor fixups for README, website, and PyPI landing page

v0.1.0#

  • First public version

Navigation