Merge branch 'develop' into serialInterval

.github/workflows/build.yaml

@@ -0,0 +1,26 @@
+name: Covasim CI workflow
+on:
+  push:
+    branches:
+    - master
+
+jobs:
+  build_and_push_docker:
+    runs-on: ubuntu-latest
+    name: Build and Publish Docker Image
+    steps:
+      - name: Checkout sources
+        uses: actions/checkout@v1
+
+      - name: Set lowercase repo env variable
+        shell: python
+        run: print("::set-env name=LOWER_REPO::{}".format('${{github.repository}}'.lower()))
+      - name: Build Docker
+        run: |
+          docker build -f docker/Dockerfile -t docker.pkg.github.com/${LOWER_REPO}/covasim:latest .
+      - name: Push the image to GPR
+        run: |
+          docker login docker.pkg.github.com -u publisher -p "${GITHUB_PACKAGE_REGISTRY_TOKEN}"
+          docker push docker.pkg.github.com/${LOWER_REPO}/covasim:latest
+        env:
+          GITHUB_PACKAGE_REGISTRY_TOKEN: ${{ secrets.GITHUB_TOKEN }}

.github/workflows/ubuntu.yaml

@@ -44,4 +44,4 @@ jobs:
         run: pytest test*.py --durations=0 # Run actual tests
       - name: Run other examples
         working-directory: ./
-        run: pytest ./examples/*.py ./covasim/cruise_ship/test*.py
+        run: pytest ./examples/*.py ./covasim/cruise_ship/test*.py

.gitignore

@@ -7,7 +7,7 @@ covasim/cova_seattle/capacity_analysis
 covasim/cova_seattle/school_analysis
 covasim/cova_seattle/seattle_analysis
 results/
-
+data/epi_data
 
 # Byte-compiled / optimized / DLL files
 __pycache__/

README.md

@@ -5,6 +5,24 @@ Covasim is a stochastic agent-based simulator designed to be used for COVID-19
 indicators such as numbers of infections and peak hospital demand. Covasim can
 also be used to explore the potential impact of different interventions.
 
+<!--ts-->
+* [Requirements](#Requirements)
+* [Quick start guide](#quick-start-guide)
+* [Detailed installation instructions](#detailed-installation-instructions)
+* [Detailed usage](#detailed-usage)
+* [Module structure](#module-structure)
+  * [cruise_ship](#cruise_ship)
+  * [webapp](#webapp)
+* [Other folders](#other-folders)
+	* [bin](#bin)
+	* [docker](#docker)
+	* [examples](#examples)
+	* [licenses](#licenses)
+	* [tests](#tests)
+	* [sweep](#sweep)
+* [Disclaimer](#disclaimer)
+<!--te-->
+
 
 ## Requirements
 
@@ -70,7 +88,7 @@ follows:
   This shows a more complex example, including running an intervention scenario, plotting uncertainty, and performing a health systems analysis.
 
 
-## Structure
+## Module structure
 
 All core model code is located in the `covasim` subfolder; standard usage is
 `import covasim as cv`. The other subfolders, `cruise_ship` and `webapp`, are
@@ -80,9 +98,6 @@ The model consists of two core classes: the `Person` class (which contains
 information on health state), and the `Sim` class (which contains methods for
 running, calculating results, plotting, etc.).
 
-
-### covasim
-
 The structure of the `covasim` folder is as follows:
 
 * `base.py`: The `ParsObj` class, plus basic methods of the `BaseSim` class, and associated functions.
@@ -107,14 +122,14 @@ slight variations to the model (`model.py`).
 
 ### webapp
 
-For running the interactive web application: please see the `README.md` in that
-folder for more information.
+For running the interactive web application. See the [webapp readme](./covasim/webapp) for more information.
 
 
 ## Other folders
 
 Please see the readme in each subfolder for more information.
 
+
 ### bin
 
 This folder contains a command-line interface (CLI) version of Covasim; example usage:
@@ -126,7 +141,12 @@ covasim --pars "{pop_size:20000, pop_infected:1, n_days:360, rand_seed:1}"
 Note: the CLI is currently not compatible with Windows. You will need to add
 this folder to your path to run from other folders.
 
-### docker
+### data
+
+Scripts to automatically scrape data (including demographics and COVID epidemiology data),
+and the data files themselves (which are not part of the repository).
+
+###  docker
 
 This folder contains the `Dockerfile` and other files that allow Covasim to be
 run as a webapp via Docker.
@@ -143,6 +163,10 @@ Licensing information and legal notices.
 
 Integration, development, and unit tests.
 
+### sweep
+
+Utilites for hyperparameter sweeps, using [Weights and Biases](https://www.wandb.com/). See the [sweep readme](./sweep) for more information.
+
 
 ## Disclaimer
 

covasim/README.md

@@ -6,7 +6,8 @@ This file describes the expected behavior of each parameter in the model. Note:
 * `pop_scale`    = Multiplicative scale for results. Test: run 2 sims, set to 10, `sim.results['cum_exposed']` in 2nd sim should be 10x higher than first.
 * `pop_size`     = Nmber of people in the simulation. Test: `len(sim.people)` should equal this number.
 * `pop_infected` = Initial number of people infected. Test: if 0, there should be no infections; if equals `n`, should be no _new_ infections.
-* `pop_type`   = Whether or not to use the `synthpops` library for contact matrices. Consult that library's documentation for tests.
+* `pop_type`   = The type of population structure to use with the model; default `'random'`, using Seattle demographics; other options are `'hybrid'` (structured households and random school, work, and community contacts), `'clustered'` (all clustered contacts), and `'synthpops'` (requires the `synthpops` library, but uses data-based population structure)
+* `location`   = Which country or state to load demographic data from (optional)
 
 ## Simulation parameters
 * `start_day`    = The calendar date of the start day of the simulation.

covasim/__init__.py

@@ -2,6 +2,7 @@
 from .version import __version__, __versiondate__, __license__
 print(__license__)
 
+
 #%% Check that requirements are met
 from . import requirements
 

covasim/base.py

@@ -146,20 +146,25 @@ def set_seed(self, seed=-1):
     def n(self):
         ''' Count the number of people -- if it fails, assume none '''
         try: # By default, the length of the people dict
-            output = len(self.people)
+            return len(self.people)
         except: # If it's None or missing
-            output = 0
-        return output
+            return 0
 
     @property
     def npts(self):
         ''' Count the number of time points '''
-        return int(self['n_days'] + 1)
+        try:
+            return int(self['n_days'] + 1)
+        except:
+            return 0
 
     @property
     def tvec(self):
         ''' Create a time vector '''
-        return np.arange(self.npts)
+        try:
+            return np.arange(self.npts)
+        except:
+            return np.arange([])
 
     @property
     def datevec(self):
@@ -171,7 +176,10 @@ def datevec(self):
             simulation time step
 
         """
-        return self['start_day'] + self.tvec * dt.timedelta(days=1)
+        try:
+            return self['start_day'] + self.tvec * dt.timedelta(days=1)
+        except:
+            return []
 
 
     def inds2dates(self, inds, dateformat=None):

covasim/data/__init__.py

@@ -0,0 +1 @@
+from .loaders import *

covasim/data/loaders.py

@@ -0,0 +1,93 @@
+'''
+Load data
+'''
+
+#%% Housekeeping
+import numpy as np
+import sciris as sc
+from . import country_age_distributions as cad
+
+__all__ = ['get_age_distribution']
+
+
+def get_age_distribution(location=None):
+    '''
+    Load age distribution for a given country or countries.
+
+    Args:
+        location (str or list): name of the country or countries to load the age distribution for
+
+    Returns:
+        age_data (array): Numpy array of age distributions
+    '''
+
+    # Load the raw data
+    json = cad.get_country_age_distributions()
+    countries = [entry["country"].lower() for entry in json] # Pull out available countries
+
+    # Set parameters
+    max_age = 99
+    if location is None:
+        location = countries
+    else:
+        location = sc.promotetolist(location)
+
+    # Define a mapping for common mistakes
+    mapping = {
+       'Bolivia':        'Bolivia (Plurinational State of)',
+       'Burkina':        'Burkina Faso',
+       'Cape Verde':     'Cabo Verdeo',
+       'Hong Kong':      'China, Hong Kong Special Administrative Region',
+       'Macao':          'China, Macao Special Administrative Region',
+       "Cote d'Ivore":   'Côte d’Ivoire',
+       'DRC':            'Democratic Republic of the Congo',
+       'Iran':           'Iran (Islamic Republic of)',
+       'Laos':           "Lao People's Democratic Republic",
+       'Micronesia':     'Micronesia (Federated States of)',
+       'Korea':          'Republic of Korea',
+       'South Korea':    'Republic of Korea',
+       'Moldova':        'Republic of Moldova',
+       'Russia':         'Russian Federation',
+       'Palestine':      'State of Palestine',
+       'Syria':          'Syrian Arab Republic',
+       'Taiwan':         'Taiwan Province of China',
+       'Macedonia':      'The former Yugoslav Republic of Macedonia',
+       'UK':             'United Kingdom of Great Britain and Northern Ireland',
+       'United Kingdom': 'United Kingdom of Great Britain and Northern Ireland',
+       'Tanzania':       'United Republic of Tanzania',
+       'USA':            'United States of America',
+       'United States':  'United States of America',
+       'Venezuela':      'Venezuela (Bolivarian Republic of)',
+       'Vietnam':        'Viet Nam',
+        }
+    mapping = {key.lower():val.lower() for key,val in mapping.items()} # Convert to lowercase
+
+    result = {}
+    for loc in location:
+        loc = loc.lower()
+        if loc in mapping:
+            loc = mapping[loc]
+        try:
+            ind = countries.index(loc.lower())
+            entry = json[ind]
+        except ValueError:
+            suggestions = sc.suggest(loc, countries, n=4)
+            errormsg = f'Location "{loc}" not recognized, did you mean {suggestions}?'
+            raise ValueError(errormsg)
+        age_distribution = entry["ageDistribution"]
+        total_pop = sum(age_distribution.values())
+        local_pop = []
+
+        for age, age_pop in age_distribution.items():
+            if age[-1] == '+':
+                val = [int(age[:-1]), max_age, age_pop/total_pop]
+            else:
+                ages = age.split('-')
+                val = [int(ages[0]), int(ages[1]), age_pop/total_pop]
+            local_pop.append(val)
+        result[loc] = np.array(local_pop)
+
+    if len(location) == 1:
+        result = result[loc]
+
+    return result

covasim/defaults.py

@@ -17,7 +17,7 @@
         'severe':      'Number of severe cases',
         'critical':    'Number of critical cases',
         'diagnosed':   'Number of confirmed cases',
-        'quarantined': 'Number in quarantined',
+        'quarantined': 'Number in quarantine',
 }
 
 # The types of result that are counted as flows -- used in sim.py; value is the label suffix

covasim/parameters.py

@@ -26,7 +26,8 @@ def make_pars(set_prognoses=False, prog_by_age=True, use_layers=False, **kwargs)
     # Population parameters
     pars['pop_size']     = 20e3 # Number ultimately susceptible to CoV
     pars['pop_infected'] = 10 # Number of initial infections
-    pars['pop_type']     = 'random' # What type of population data to use -- random (fastest), synthpops (best), realistic (compromise), or clustered (not recommended)
+    pars['pop_type']     = 'random' # What type of population data to use -- random (fastest), synthpops (best), hybrid (compromise), or clustered (not recommended)
+    pars['location']     = None # What location to load data from -- default Seattle
 
     # Simulation parameters
     pars['start_day']  = '2020-03-01' # Start day of the simulation