Release 0.10.8

.github/workflows/ci_dependencies.sh

@@ -47,9 +47,7 @@ do
       pip install \
         git+https://github.com/amici-dev/amici.git@develop#egg=amici\&subdirectory=python/sdist
       # install test suite
-      git clone --depth 1 \
-        https://github.com/petab-dev/petab_test_suite .tmp/petab_test_suite
-      pip install -e .tmp/petab_test_suite
+      pip install petabtests
     ;;
 
     docs)

.github/workflows/ci_push.yml

@@ -8,7 +8,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: [3.8]
+        python-version: [3.8, 3.7]
 
     steps:
     - name: Check out repository
@@ -29,7 +29,7 @@ jobs:
     - name: Install dependencies
       run: |
         .github/workflows/ci_dependencies.sh base R
-        pip install -e .[R,test,code_quality]
+        pip install -e .[R,test]
 
     - name: Run tests
       timeout-minutes: 10
@@ -47,7 +47,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: [3.8]
+        python-version: [3.8, 3.7]
 
     steps:
     - name: Check out repository
@@ -85,7 +85,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: [3.8]
+        python-version: [3.8, 3.7]
 
     steps:
     - name: Check out repository
@@ -123,7 +123,7 @@ jobs:
     runs-on: macos-latest
     strategy:
       matrix:
-        python-version: [3.8]
+        python-version: [3.8, 3.7]
 
     steps:
       - name: Check out repository
@@ -147,7 +147,7 @@ jobs:
           pip install -e .[test]
 
       - name: Run tests
-        timeout-minutes: 3
+        timeout-minutes: 4
         run: |
           python -m pytest --cov=pyabc --cov-report=xml test/base/test_macos.py
 
@@ -184,7 +184,7 @@ jobs:
         pip install -e .[quality]
 
     - name: Run flake8
-      timeout-minutes: 1
+      timeout-minutes: 2
       run: ./run_flake8.sh
 
     - name: Run pyroma

doc/examples/conversion_reaction.ipynb

@@ -19,6 +19,17 @@
     ":download:`Ordinary Differential Equations <conversion_reaction.ipynb>`."
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "<div class=\"alert alert-warning\">\n",
+    "\n",
+    "**Note:** Before you use pyABC to parametrize your ODE, please be aware of potential errors introduced by inadequately representing the data generation process, see also the \"Measurement noise assessment\" notebook. For deterministic models, there are often more efficient alternatives to ABC, check out for example our tool <a href=\"https://github.com/ICB-DCM/pyPESTO\">pyPESTO</a>.\n",
+    "\n",
+    "</div>"
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {
@@ -456,7 +467,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.7.4"
+   "version": "3.8.5"
   }
  },
  "nbformat": 4,

doc/index.rst

@@ -47,35 +47,11 @@ If you use it in your work, you can cite the paper:
    :caption: User's guide
 
    what
-
-.. toctree::
-   :maxdepth: 2
-
    installation
-
-.. toctree::
-   :maxdepth: 2
-
    examples
-
-.. toctree::
-   :maxdepth: 2
-
    sampler
-
-.. toctree::
-   :maxdepth: 2
-
    sge
-
-.. toctree::
-   :maxdepth: 2
-
-   export_db
-
-.. toctree::
-   :maxdepth: 2
-
+   export_db 
    web_visualization
 
 
@@ -84,23 +60,15 @@ If you use it in your work, you can cite the paper:
    :caption: About
 
    releasenotes
-
-.. toctree::
-   :maxdepth: 2
-
    about
+   references
 
 
 .. toctree::
    :maxdepth: 2
    :caption: Developer's guide
 
    contribute
-
-
-.. toctree::
-   :maxdepth: 2
-
    deploy
 
 

doc/references.rst

@@ -0,0 +1,36 @@
+.. _references:
+
+References
+==========
+
+List of publications using pyABC, so far in total amounting to 16 citations.
+If you applied pyABC in your work, please consider notifying us, so we can include your work here.
+
+2020
+----
+
+- Schälte, Yannik & Hasenauer, Jan. (2020). Efficient exact inference for dynamical systems with noisy measurements using sequential approximate Bayesian computation. Bioinformatics (Oxford, England). 36. i551-i559. `https://doi.org/10.1093/bioinformatics/btaa397`. 
+- Ruske, Liam & Kursawe, Jochen & Tsakiridis, Anestis & Wilson, Valerie & Fletcher, Alexander & Blythe, Richard & Schumacher, Linus. (2020). Coupled differentiation and division of embryonic stem cells inferred from clonal snapshots. Physical Biology. `https://doi.org/10.1088/1478-3975/aba041`. 
+- Houston, Charles & Marchand, B. & Engelbert, L. & Cantwell, Chris. (2020). Reducing complexity and unidentifiability when modelling human atrial cells. Philosophical Transactions of The Royal Society A Mathematical Physical and Engineering Sciences. 378. 20190339. `https://doi.org/10.1098/rsta.2019.0339`. 
+- Sid Ahmed, Samy & Bundgaard, Nils & Graw, Frederik & Fackler, Oliver. (2020). Environmental Restrictions: A New Concept Governing HIV-1 Spread Emerging from Integrated Experimental-Computational Analysis of Tissue-Like 3D Cultures. Cells. 9. 1112. `https://doi.org/10.3390/cells9051112`. 
+- Sendera, Marcin & Duane, Gregory & Dzwinel, Witold. (2020). Supermodeling: The Next Level of Abstraction in the Use of Data Assimilation. `https://doi.org/10.1007/978-3-030-50433-5_11`. 
+- Lupperger, Valerio & Marr, Carsten & Chapouton, Prisca. (2020). Aggregated spatio-temporal division patterns emerge from reoccurring divisions of neural stem cells. `https://doi.org/10.1101/2020.03.20.999748`. 
+- Bheda, Poonam & Aguilar-Gómez, Diana & Becker, Nils & Becker, Johannes & Stavrou, Emmanouil & Kukhtevich, Igor & Höfer, Thomas & Maerkl, Sebastian & Charvin, Gilles & Marr, Carsten & Kirmizis, Antonis & Schneider, Robert. (2020). Single-Cell Tracing Dissects Regulation of Maintenance and Inheritance of Transcriptional Reinduction Memory. Molecular Cell. 78. `https://doi.org/10.1016/j.molcel.2020.04.016`. 
+- Nadjahi, Kimia & Bortoli, Valentin & Durmus, Alain & Badeau, Roland & Simsekli, Umut. (2020). Approximate Bayesian Computation with the Sliced-Wasserstein Distance. 5470-5474. `https://doi.org/10.1109/ICASSP40776.2020.9054735`. 
+
+2019
+----
+
+- Bölts, Jan & Lueckmann, Jan-Matthis & Goncalves, Pedro & Sprekeler, Henning & Macke, Jakob. (2019). Comparing neural simulations by neural density estimation. `https://doi.org/10.32470/CCN.2019.1291-0`.
+- Raynal, Louis & Marin, Jean-Michel & Pudlo, Pierre & Ribatet, Mathieu & Robert, Christian & Estoup, Arnaud. (2019). Genetics and population analysis ABC random forests for Bayesian parameter inference. Bioinformatics. 35. 1720. `https://doi.org/0.1093/bioinformatics/bty867`
+- Vu, Hanh & Mansour, Sarah & Kücken, Michael & Blasse, Corinna & Basquin, Cyril & Azimzadeh, Juliette & Myers, Eugene & Brusch, Lutz & Rink, Jochen. (2019). Dynamic Polarization of the Multiciliated Planarian Epidermis between Body Plan Landmarks. Developmental Cell. 51. 526-542.e6. `https://doi.org/10.1016/j.devcel.2019.10.022`. 
+- Imle, Andrea & Kumberger, Peter & Schnellbächer, Nikolas & Fehr, Jana & Carrillo-Bustamante, Paola & Ales, Janez & Schmidt, Philip & Ritter, Christian & Godinez, William & Müller, Barbara & Rohr, Karl & Hamprecht, Fred & Schwarz, Ulrich & Graw, Frederik & Fackler, Oliver. (2019). Experimental and computational analyses reveal that environmental restrictions shape HIV-1 spread in 3D cultures. Nature Communications. 10. `https://doi.org/10.1038/s41467-019-09879-3`. 
+- Raynal, Louis. Bayesian statistical inference for intractable likelihood models. Diss. Université Montpellier, 2019.
+- Kumberger, Peter Johannes Manuel Eduard. Quantifying the Impact of Cell-to-Cell Transmission on Viral Spread. Diss. 2019.
+
+2018
+----
+
+- Cantwell, Chris & Mohamied, Yumnah & Tzortzis, Konstantinos & Garasto, Stef & Houston, Charles & Chowdhury, Rasheda & Ng, Fu Siong & Bharath, Anil & Peters, Nicholas. (2019). Rethinking multiscale cardiac electrophysiology with machine learning and predictive modelling. Computers in Biology and Medicine. 104. 339-351. `https://doi.org/10.1016/j.compbiomed.2018.10.015`. 
+- Klinger, Emmanuel G. Approximate Bayesian Model Selection for Local Cortical Networks at Cellular Resolution. Diss. Technische Universität München, 2018.
+

doc/releasenotes.rst

@@ -8,6 +8,17 @@ Release Notes
 ...........
 
 
+0.10.8 (2020-11-27)
+-------------------
+
+* Allow to normalize acceptance rate plots by ESS (#346).
+* Add a list of pyABC references (#348).
+* Update to petabtests 0.0.0a5 (#362).
+* Add stopping criterion for total number of samples (#364).
+* Remove dill dependency, thus fixing a cloudpickle error, and
+  run selected tests also with python 3.7 (#367).
+
+
 0.10.7 (2020-08-20)
 -------------------
 

pyabc/inference/smc.py

@@ -539,25 +539,24 @@ def _create_transition_pdf(self, t: int, transitions):
     def run(self,
             minimum_epsilon: float = None,
             max_nr_populations: int = np.inf,
-            min_acceptance_rate: float = 0.) -> History:
+            min_acceptance_rate: float = 0.,
+            max_total_nr_simulations: int = np.inf) -> History:
         """
         Run the ABCSMC model selection until either of the stopping
         criteria is met.
 
         Parameters
         ----------
-
-        minimum_epsilon: float, optional
+        minimum_epsilon:
             Stop if epsilon is smaller than minimum epsilon specified here.
             Defaults in general to 0.0, and to 1.0 for a Temperature epsilon.
-
-        max_nr_populations: int, optional (default = np.inf)
+        max_nr_populations:
             The maximum number of populations. Stop if this number is reached.
-
-        min_acceptance_rate: float, optional (default = 0.0)
+        min_acceptance_rate:
             Minimal allowed acceptance rate. Sampling stops if a population
             has a lower rate.
-
+        max_total_nr_simulations:
+            Bound on the total number of evaluations.
 
         Population after population is sampled and particles which are close
         enough to the observed data are accepted and added to the next
@@ -672,6 +671,9 @@ def run(self,
             elif acceptance_rate < min_acceptance_rate:
                 logger.info("Stopping: minimum acceptance rate.")
                 break
+            elif self.history.total_nr_simulations >= max_total_nr_simulations:
+                logger.info("Stopping: total simulations budget.")
+                break
 
             # increment t
             t += 1

pyabc/sampler/mapping.py

@@ -1,7 +1,7 @@
 import functools
 import random
 
-import dill as pickle
+import cloudpickle as pickle
 import numpy as np
 
 from .base import Sampler

pyabc/version.py

@@ -1 +1 @@
-__version__ = "0.10.7"
+__version__ = "0.10.8"

pyabc/visualization/sample.py

@@ -5,6 +5,7 @@
 
 from ..storage import History
 from .util import to_lists_or_default
+from ..weighted_statistics import effective_sample_size
 
 
 def plot_sample_numbers(
@@ -263,7 +264,8 @@ def plot_acceptance_rates_trajectory(
         yscale: str = 'lin',
         size: tuple = None,
         ax: mpl.axes.Axes = None,
-        colors: List[str] = None):
+        colors: List[str] = None,
+        normalize_by_ess: bool = False):
     """
     Plot of acceptance rates over all iterations, i.e. one trajectory
     per history.
@@ -288,6 +290,9 @@ def plot_acceptance_rates_trajectory(
         The size of the plot in inches.
     ax: matplotlib.axes.Axes, optional
         The axis object to use.
+    normalize_by_ess: bool, optional (default = False)
+        Indicator to use effective sample size for the acceptance rate in
+        place of the population size.
 
     Returns
     -------
@@ -312,9 +317,16 @@ def plot_acceptance_rates_trajectory(
         # note: the first entry of time -1 is trivial and is thus ignored here
         h_info = history.get_all_populations()
         times.append(np.array(h_info['t'])[1:])
+        if normalize_by_ess:
+            ess = np.zeros(len(h_info['t']) - 1)
+            for t in np.array(h_info['t'])[1:]:
+                w = history.get_weighted_distances(t=t)['w']
+                ess[t-1] = effective_sample_size(w)
+            pop_sizes.append(ess)
+        else:
+            pop_sizes.append(np.array(
+                history.get_nr_particles_per_population().values[1:]))
         samples.append(np.array(h_info['samples'])[1:])
-        pop_sizes.append(np.array(
-            history.get_nr_particles_per_population().values[1:]))
 
     # compute acceptance rates
     rates = []

run_flake8.sh

@@ -1,3 +1,5 @@
 #!/bin/sh
 
-python3 -m flake8 pyabc test --extend-ignore='S403,S301' --per-file-ignores='*/__init__.py:F401 test/*:T001,S101'
+python3 -m flake8 pyabc test \
+  --extend-ignore='S403,S301,C408' \
+  --per-file-ignores='*/__init__.py:F401 test/*:T001,S101'

setup.cfg

@@ -53,7 +53,6 @@ install_requires =
     flask >= 1.1.2
     bokeh >= 2.1.1
     redis >= 2.10.6
-    dill >= 0.3.2
     gitpython >= 3.1.7
     scikit-learn >= 0.23.1
     matplotlib >= 3.3.0
@@ -77,7 +76,7 @@ packages = find:
 
 [options.extras_require]
 R =
-    rpy2 >= 3.2.0
+    rpy2 >= 3.3.6
     cffi >= 1.13.1
 petab =
     petab >= 0.1.8

test/base/test_stop_sampling.py

@@ -9,14 +9,17 @@
 
 
 def model(x):
+    """Some model"""
     return {"par": x["par"] + np.random.randn()}
 
 
 def dist(x, y):
+    """Some distance"""
     return abs(x["par"] - y["par"])
 
 
 def test_stop_acceptance_rate_too_low(db_path):
+    """Test the acceptance rate condition."""
     abc = ABCSMC(model, Distribution(par=st.uniform(0, 10)), dist, pop_size)
     abc.new(db_path, {"par": .5})
     history = abc.run(-1, 8, min_acceptance_rate=set_acc_rate)
@@ -28,6 +31,7 @@ def test_stop_acceptance_rate_too_low(db_path):
 
 
 def test_stop_early(db_path):
+    """Test early stopping inside a generation."""
     mc_sampler = MulticoreEvalParallelSampler(check_max_eval=True)
     sc_sampler = SingleCoreSampler(check_max_eval=True)
     for sampler in [mc_sampler, sc_sampler]:
@@ -44,3 +48,18 @@ def test_stop_early(db_path):
             df["particles"] / (df["samples"] - (n_procs-1))
 
         assert df["corrected_acceptance_rate"].iloc[-1] >= set_acc_rate
+
+
+def test_total_nr_simulations(db_path):
+    """Test the total number of samples condition."""
+    abc = ABCSMC(model, Distribution(par=st.uniform(0, 10)), dist, pop_size)
+    abc.new(db_path, {"par": .5})
+    max_total_nr_sim = 142
+    history = abc.run(-1, 100, max_total_nr_simulations=max_total_nr_sim)
+    assert history.total_nr_simulations >= max_total_nr_sim
+    # Directly check on the history
+    df = history.get_all_populations()
+    # Make sure budget is not exceeded yet in previous iteration
+    assert sum(df['samples'][:-1]) < max_total_nr_sim
+    # Just to make sure .total_nr_simulations does what it's supposed to
+    assert sum(df['samples']) == history.total_nr_simulations