hpvsim 2.2.7__tar.gz → 2.3.0__tar.gz

{hpvsim-2.2.7 → hpvsim-2.3.0}/.gitignore

@@ -8,6 +8,9 @@ _inv/
 api/
 .jupyter_cache/
 
+# Regression baselines (regenerate via tests/generate_v2_baselines.py)
+tests/regression_baselines/
+
 # Other files
 .idea
 .vscode

{hpvsim-2.2.7 → hpvsim-2.3.0}/CHANGELOG.md

@@ -3,10 +3,27 @@ that may result in differences in model output, or are required in order
 to run an old parameter set with the current version, are flagged with
 the term "Regression information".
 
+## Version 2.3.0 (2026-04-20)
+
+- Fixes dt-dependent results by scaling partnership formation rates to
+  per-timestep probabilities; `layer_probs` and cross-layer defaults
+  converted to annual probabilities.
+- *Regression information*: If workflows from v2.2.6 or earlier override default `layer_probs`, `f_cross_layer`, or `m_cross_layer` values
+  and have timesteps not equal to 1 year, then the probabilities must be converted to annual probabilities instead of per-timestep probabilities using this formula: `1 - (1 - prob) ** dt`
+- *Regression information:* baseline model outputs change; baselines have been regenerated.
+- Fixes `precins` flow never being incremented; removes redundant `dysplasias` flow (was an alias of `cins`).
+- Adds test coverage for previously untested code paths.
+- Vaccine immunity is now sterilizing (all-or-nothing) rather than leaky (per-contact). `imm_init` sets the probability of sterilizing immunity; non-sterilizing recipients get leaky protection at the `imm_init` level. Default is 0.95.
+- Adds per-timestep transmission logging (`sim._transmission_log`) for downstream analysis of transmission chains.
+- Calibration now supports resuming from an existing database via `keep_db=True`, running only the remaining trials.
+- Calibration workers catch exceptions instead of crashing the entire run.
+- Fixes `res_to_plot` indexing bug in `Calibration.plot()`.
+- *Regression information*: vaccine efficacy will differ from previous versions due to the immunity model change.
+
 ## Version 2.2.7 (2026-04-22)
 
 - Fix cancer treatment results always being blank: `BaseTreatment.check_eligibility` was excluding cancer patients (preventing radiation from running), `BaseTreatment.apply` was writing to CIN fields for cancer treatment, and `cum_cancer_treated` cumsum used the wrong source array
-- *Github info* PR [92](https://github.com/starsimhub/hpvsim/pull/92), issue [91](https://github.com/starsimhub/hpvsim/issues/91)
+- *Github info* PR [94](https://github.com/starsimhub/hpvsim/pull/94), issue [91](https://github.com/starsimhub/hpvsim/issues/91)
 
 ## Version 2.2.6 (2026-04-17)
 

{hpvsim-2.2.7/hpvsim.egg-info → hpvsim-2.3.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: hpvsim
-Version: 2.2.7
+Version: 2.3.0
 Summary: HPVsim: Human Papillomavirus Simulator
 Author: Robyn Stuart, Jamie Cohen, Cliff Kerr, Romesh Abeysuriya, Mariah Boudreau, Daniel Klein, Hao Hu
 Maintainer-email: HPVsim Team <info@hpvsim.org>

{hpvsim-2.2.7 → hpvsim-2.3.0}/hpvsim/analysis.py

@@ -742,7 +742,6 @@ class age_results(Analyzer):
         mapping = {
             'infections': ['date_exposed', 'infectious'],
             'cin':  ['date_cin', 'cin'],
-            'dysplasias':  ['date_cin', 'cin'],
             'cins':  ['date_cin', 'cin'],
             'cancers': ['date_cancerous', 'cancerous'],
             'cancer': ['date_cancerous', 'cancerous'],

{hpvsim-2.2.7 → hpvsim-2.3.0}/hpvsim/calibration.py

@@ -89,8 +89,10 @@ class Calibration(sc.prettyobj):
         if keep_db   is None: keep_db   = False
         if storage   is None: storage   = f'sqlite:///{db_name}'
         if total_trials is not None: n_trials = int(np.ceil(total_trials/n_workers))
-        self.run_args   = sc.objdict(n_trials=int(n_trials), n_workers=int(n_workers), name=name, db_name=db_name,
-                                     keep_db=keep_db, storage=storage, rand_seed=rand_seed, sampler=sampler)
+        total_trials = int(n_trials * n_workers)
+        self.run_args   = sc.objdict(n_trials=int(n_trials), n_workers=int(n_workers), total_trials=total_trials,
+                                     name=name, db_name=db_name, keep_db=keep_db, storage=storage,
+                                     rand_seed=rand_seed, sampler=sampler)
 
         # Handle other inputs
         self.label          = label
@@ -412,7 +414,11 @@ class Calibration(sc.prettyobj):
         else:
             op.logging.set_verbosity(op.logging.ERROR)
         study = op.load_study(storage=self.run_args.storage, study_name=self.run_args.name, sampler = self.run_args.sampler)
-        output = study.optimize(self.run_trial, n_trials=self.run_args.n_trials, callbacks=None)
+        try:
+            output = study.optimize(self.run_trial, n_trials=self.run_args.n_trials, callbacks=None)
+        except Exception as E:
+            print(f'Worker failed with error: {E}')
+            output = None
         return output
 
 
@@ -438,9 +444,12 @@ class Calibration(sc.prettyobj):
             print('Could not delete study, skipping...')
             print(str(E))
         if os.path.exists(self.run_args.db_name):
-            os.remove(self.run_args.db_name)
-            if self.verbose:
-                print(f'Removed existing calibration {self.run_args.db_name}')
+            try:
+                os.remove(self.run_args.db_name)
+                if self.verbose:
+                    print(f'Removed existing calibration {self.run_args.db_name}')
+            except PermissionError:
+                print(f'Could not remove {self.run_args.db_name} (file in use), skipping...')
         return
 
 
@@ -455,7 +464,8 @@ class Calibration(sc.prettyobj):
             raise NotImplementedError('Implemented but does not work')
         else:
             sampler = None
-        output = op.create_study(storage=self.run_args.storage, study_name=self.run_args.name, sampler=sampler)
+        output = op.create_study(storage=self.run_args.storage, study_name=self.run_args.name, sampler=sampler,
+                                 load_if_exists=self.run_args.keep_db)
         return output
 
 
@@ -485,7 +495,23 @@ class Calibration(sc.prettyobj):
         # Run the optimization
         t0 = sc.tic()
         self.make_study()
-        self.run_workers()
+
+        # If resuming with keep_db, check for existing trials and only run the remainder
+        if self.run_args.keep_db:
+            study = op.load_study(storage=self.run_args.storage, study_name=self.run_args.name)
+            n_existing = len([t for t in study.trials if t.state == op.trial.TrialState.COMPLETE])
+            if n_existing > 0:
+                n_remaining = max(0, self.run_args.total_trials - n_existing)
+                if n_remaining == 0:
+                    print(f'Calibration already has {n_existing} completed trials, skipping workers')
+                else:
+                    self.run_args.n_trials = int(np.ceil(n_remaining / self.run_args.n_workers))
+                    print(f'Resuming calibration: {n_existing} trials complete, running ~{n_remaining} more')
+                    self.run_workers()
+            else:
+                self.run_workers()
+        else:
+            self.run_workers()
         study = op.load_study(storage=self.run_args.storage, study_name=self.run_args.name, sampler = self.run_args.sampler)
         self.best_pars = sc.objdict(study.best_params)
         self.elapsed = sc.toc(t0, output=True)
@@ -500,6 +526,7 @@ class Calibration(sc.prettyobj):
         self.analyzer_results = []
         self.sim_results = []
         self.extra_sim_results = []
+        loaded_trials = set()
         if load:
             print('Loading saved results...')
             for trial in study.trials:
@@ -510,6 +537,7 @@ class Calibration(sc.prettyobj):
                     self.sim_results.append(results['sim'])
                     self.analyzer_results.append(results['analyzer'])
                     self.extra_sim_results.append(results['extra_sim_results'])
+                    loaded_trials.add(n)
                     if tidyup:
                         try:
                             os.remove(filename)
@@ -524,7 +552,7 @@ class Calibration(sc.prettyobj):
 
         # Compare the results
         self.initial_pars, self.par_bounds = self.sim_to_sample_pars()
-        self.parse_study(study)
+        self.parse_study(study, loaded_trials=loaded_trials)
 
         # Tidy up
         self.calibrated = True
@@ -534,7 +562,7 @@ class Calibration(sc.prettyobj):
         return self
 
 
-    def parse_study(self, study):
+    def parse_study(self, study, loaded_trials=None):
         '''Parse the study into a data frame -- called automatically '''
         best = study.best_params
         self.best_pars = best
@@ -549,6 +577,8 @@ class Calibration(sc.prettyobj):
                 data[key] = val
             if data['mismatch'] is None:
                 failed_trials.append(data['index'])
+            elif loaded_trials is not None and trial.number not in loaded_trials:
+                failed_trials.append(data['index'])
             else:
                 results.append(data)
         print(f'Processed {n_trials} trials; {len(failed_trials)} failed')
@@ -653,7 +683,7 @@ class Calibration(sc.prettyobj):
 
         # determine how many results to plot
         if res_to_plot is not None:
-            index_to_plot = self.df.iloc[0:res_to_plot, 0].values
+            index_to_plot = self.df.index[0:res_to_plot].values
             analyzer_results = [analyzer_results[i] for i in index_to_plot]
             sim_results = [sim_results[i] for i in index_to_plot]
 

{hpvsim-2.2.7 → hpvsim-2.3.0}/hpvsim/defaults.py

@@ -271,6 +271,7 @@ class Flow():
 
 flows = [
     Flow('infections',              color='#c78f65',    label='Infections'),
+    Flow('precins',                 color='#c1ad71',    label='Pre-CINs'),
     Flow('cins',                    color='#b86113',    label='CINs'),
     Flow('cancers',                 color='#5f5cd2',    label='Cancers'),
     Flow('cancer_deaths',           color='#000000',    label='Cancer deaths', by_genotype=False),

{hpvsim-2.2.7 → hpvsim-2.3.0}/hpvsim/interventions.py

@@ -1414,22 +1414,52 @@ class tx(Product):
 
 
 class vx(Product):
-    ''' Vaccine product '''
+    '''
+    Vaccine product.
+
+    imm_init controls the probability that a vaccinated person gets sterilizing
+    (all-or-nothing) immunity. People who don't get sterilizing immunity still
+    receive leaky (per-contact) protection equal to imm_init.
+
+    imm_init can be:
+        - a float (e.g. 0.95): used directly as the sterilizing probability
+        - a dict with dist/par keys (legacy beta-distribution format): the
+          mean of the distribution is used as the sterilizing probability
+    '''
     def __init__(self, genotype_pars=None, imm_init=None, imm_boost=None):
         self.genotype_pars = genotype_pars
-        self.imm_init = imm_init
         self.imm_boost = imm_boost
         self.imm_source = None # Set during immunity initialization. Warning, fragile!!!
         if (imm_init is None and imm_boost is None) or (imm_init is not None and imm_boost is not None):
             errormsg = 'Must provide either an initial immune effect (for first doses) or an immune boosting effect (for subsequent doses), not both/neither.'
             raise ValueError(errormsg)
 
+        # Convert imm_init to a float (sterilizing probability)
+        if imm_init is not None:
+            if isinstance(imm_init, dict):
+                if imm_init.get('dist') == 'beta':
+                    a, b = imm_init['par1'], imm_init['par2']
+                    self.imm_init = a / (a + b)
+                elif imm_init.get('dist') == 'beta_mean':
+                    self.imm_init = imm_init['par1']
+                else:
+                    # Estimate the mean of an unknown distribution via Monte Carlo;
+                    # 10k samples is just for accuracy, not related to agent count
+                    self.imm_init = float(np.mean(hpu.sample(**imm_init, size=10000)))
+            else:
+                self.imm_init = float(imm_init)
+        else:
+            self.imm_init = None
+
 
     def administer(self, people, inds):
         ''' Apply the vaccine to the requested people indices. '''
         inds = inds[people.alive[inds]]  # Skip anyone that is dead
         if self.imm_init is not None:
-            people.peak_imm[self.imm_source, inds] = hpu.sample(**self.imm_init, size=len(inds))* people.rel_imm[inds]
+            n = len(inds)
+            sterilizing = np.random.random(n) < self.imm_init
+            peak = np.where(sterilizing, 1.0, self.imm_init)
+            people.peak_imm[self.imm_source, inds] = peak * people.rel_imm[inds]
         elif self.imm_boost is not None:
             people.peak_imm[self.imm_source, inds] *= self.imm_boost
         people.t_imm_event[self.imm_source, inds] = people.t
@@ -1515,7 +1545,7 @@ def default_vx(prod_name=None):
     dfvx = pd.read_csv(datafiles.vx) # Read in dataframe with parameters
     vxprods = dict()
     for name in dfvx.name.unique():
-        vxprods[name]       = vx(genotype_pars=dfvx[dfvx.name==name], imm_init=dict(dist='beta', par1=30, par2=2))
+        vxprods[name]       = vx(genotype_pars=dfvx[dfvx.name==name], imm_init=0.95)
         vxprods[name+'2']   = vx(genotype_pars=dfvx[dfvx.name==name], imm_boost=1.2) # 2nd dose
         vxprods[name+'3']   = vx(genotype_pars=dfvx[dfvx.name==name], imm_boost=1.1) # 3rd dose
     if prod_name is not None:   return vxprods[prod_name]

{hpvsim-2.2.7 → hpvsim-2.3.0}/hpvsim/parameters.py

@@ -73,8 +73,8 @@ def make_pars(**kwargs):
     pars['add_mixing']      = None  # Mixing matrix between clusters
     pars['debut']           = dict(f=dict(dist='normal', par1=15.0, par2=2.1), # Location-specific data should be used here if possible
                                    m=dict(dist='normal', par1=17.6, par2=1.8))
-    pars['f_cross_layer']   = 0.05  # Proportion of females who have concurrent cross-layer relationships - by layer
-    pars['m_cross_layer']   = 0.30  # Proportion of males who have concurrent cross-layer relationships - by layer
+    pars['f_cross_layer']   = 0.185  # Annual probability of females having concurrent cross-layer relationships
+    pars['m_cross_layer']   = 0.760  # Annual probability of males having concurrent cross-layer relationships
     pars['f_partners']      = None  # Distribution of preferred number of concurrent sexual partners, females
     pars['m_partners']      = None  # Distribution of preferred number of concurrent sexual partners, males
     pars['acts']            = None  # The number of sexual acts for each partnership type per year
@@ -578,16 +578,14 @@ def get_mixing(network=None):
 
         layer_probs = dict(
             m=np.array([
-                [ 0,  5,    10,    15,   20,   25,   30,   35,    40,    45,    50,   55,   60,   65,   70,   75],
-                [ 0,  0,  0.01,   0.5,  0.5,  0.5,  0.5,  0.5,   0.5,   0.5,   0.5,  0.3,  0.2,  0.1, 0.05, 0.01], # Share of females of each age who are actively seeking marriage if underpartnered
-                [ 0,  0,  0.01,   0.2,  0.3,  0.5,  0.5,  0.5,   0.5,   0.5,   0.5,  0.3,  0.2,  0.1, 0.05, 0.01]] # Share of males of each age who are actively seeking marriage if underpartnered
-                # [ 1,  1,     1,     1,    1,    1,    1,    1,     1,     1,     1,    1,    1,    1,    1,    1]] # Share of males of each age who are actively seeking marriage if underpartnered
+                [ 0,    5,      10,     15,    20,    25,    30,    35,     40,     45,     50,    55,    60,    65,    70,    75],
+                [ 0,    0,  0.0394, 0.938, 0.938, 0.938, 0.938, 0.938, 0.938, 0.938, 0.938, 0.760, 0.590, 0.344, 0.185, 0.0394], # Annual prob of females seeking marriage if underpartnered
+                [ 0,    0,  0.0394, 0.590, 0.760, 0.938, 0.938, 0.938, 0.938, 0.938, 0.938, 0.760, 0.590, 0.344, 0.185, 0.0394]] # Annual prob of males seeking marriage if underpartnered
             ),
             c=np.array([
-                [ 0,  5,    10,    15,   20,   25,   30,   35,    40,    45,    50,   55,   60,   65,   70,   75],
-                [ 0,  0,   0.2,   0.6,  0.8,  0.6,  0.4,  0.4,   0.4,   0.1,  0.02, 0.02, 0.02, 0.02, 0.02, 0.02], # Share of females of each age actively seeking casual relationships if underpartnered
-                [ 0,  0,   0.2,   0.4,  0.4,  0.4,  0.4,  0.6,   0.8,   0.6,   0.2,  0.1, 0.05, 0.02, 0.02, 0.02]] # Share of males of each age actively seeking casual relationships if underpartnered
-                # [ 1,  1,     1,     1,    1,    1,    1,    1,     1,     1,     1,    1,    1,    1,    1,    1]] # Share of males of each age actively seeking casual relationships if underpartnered
+                [ 0,    5,      10,     15,    20,    25,    30,    35,     40,     45,     50,    55,    60,    65,    70,    75],
+                [ 0,    0,  0.590, 0.974, 0.998, 0.974, 0.870, 0.870, 0.870, 0.344, 0.0776, 0.0776, 0.0776, 0.0776, 0.0776, 0.0776], # Annual prob of females seeking casual relationships if underpartnered
+                [ 0,    0,  0.590, 0.870, 0.870, 0.870, 0.870, 0.974, 0.998, 0.974, 0.590, 0.344, 0.185, 0.0776, 0.0776, 0.0776]] # Annual prob of males seeking casual relationships if underpartnered
             )
         )
 
@@ -615,9 +613,9 @@ def get_mixing(network=None):
         )
         layer_probs = dict(
             a=np.array([
-                [ 0,  5,    10,    15,   20,   25,   30,   35,    40,    45,    50,   55,   60,   65,   70,   75],
-                [ 0,  0,  0.04,   0.2,  0.6,  0.8,  0.8,  0.8,  0.75,  0.65,  0.55,  0.4,  0.4,  0.4,  0.4,  0.4], # Share of females of each age who are married
-                [ 0,  0,  0.01,  0.01,  0.2,  0.6,  0.8,  0.9,  0.90,  0.90,  0.90,  0.8,  0.7,  0.6,  0.5,  0.6]] # Share of males of each age who are married
+                [ 0,    5,      10,     15,    20,    25,    30,    35,     40,     45,     50,    55,    60,    65,    70,    75],
+                [ 0,    0,  0.151, 0.590, 0.974, 0.998, 0.998, 0.998, 0.996, 0.985, 0.959, 0.870, 0.870, 0.870, 0.870, 0.870], # Annual prob of females seeking partnership if underpartnered
+                [ 0,    0,  0.0394, 0.0394, 0.590, 0.974, 0.998, 0.999, 0.999, 0.999, 0.999, 0.998, 0.992, 0.974, 0.938, 0.974]] # Annual prob of males seeking partnership if underpartnered
             ))
 
     else:
@@ -636,42 +634,42 @@ def get_vaccine_dose_pars(default=False, vaccine=None):
     pars = dict(
 
         default = dict(
-            imm_init  = dict(dist='beta', par1=30, par2=2), # Initial distribution of immunity
+            imm_init  = 0.95, # Probability of sterilizing immunity
             doses     = 1, # Number of doses for this vaccine
             interval  = None, # Interval between doses
             imm_boost=None,  # For vaccines wiht >1 dose, the factor by which each additional boost increases immunity
         ),
 
         bivalent = dict(
-            imm_init=dict(dist='beta', par1=30, par2=2),  # Initial distribution of immunity
+            imm_init=0.95,  # Probability of sterilizing immunity
             doses=1,  # Number of doses for this vaccine
             interval=None,  # Interval between doses
             imm_boost=None,  # For vaccines wiht >1 dose, the factor by which each additional boost increases immunity
         ),
 
         bivalent_2dose = dict(
-            imm_init=dict(dist='beta', par1=30, par2=2),  # Initial distribution of immunity
+            imm_init=0.95,  # Probability of sterilizing immunity
             doses=2,  # Number of doses for this vaccine
             interval=0.5,  # Interval between doses in years
             imm_boost=1.2,  # For vaccines wiht >1 dose, the factor by which each additional boost increases immunity
         ),
 
         bivalent_3dose = dict(
-            imm_init=dict(dist='beta', par1=30, par2=2),  # Initial distribution of immunity
+            imm_init=0.95,  # Probability of sterilizing immunity
             doses=3,  # Number of doses for this vaccine
             interval=[0.2, 0.5],  # Interval between doses in years
             imm_boost=[1.2, 1.1],  # Factor by which each dose increases immunity
         ),
 
         quadrivalent = dict(
-            imm_init=dict(dist='beta', par1=30, par2=2),  # Initial distribution of immunity
+            imm_init=0.95,  # Probability of sterilizing immunity
             doses=1,  # Number of doses for this vaccine
             interval=None,  # Interval between doses
             imm_boost=None,  # For vaccines wiht >1 dose, the factor by which each additional boost increases immunity
         ),
 
         nonavalent = dict(
-            imm_init=dict(dist='beta', par1=30, par2=2),  # Initial distribution of immunity
+            imm_init=0.95,  # Probability of sterilizing immunity
             doses=1,  # Number of doses for this vaccine
             interval=None,  # Interval between doses
             imm_boost=None,  # For vaccines wiht >1 dose, the factor by which each additional boost increases immunity

{hpvsim-2.2.7 → hpvsim-2.3.0}/hpvsim/people.py

@@ -458,6 +458,16 @@ class People(hpb.BasePeople):
         # Initialize
         new_pships = dict()
 
+        # Scale annual participation rates to per-timestep probabilities so that
+        # partnership formation rates are consistent regardless of dt (issue #13)
+        dt = self.dt
+        scaled_layer_probs = {lkey: lp.copy() for lkey, lp in layer_probs.items()}
+        for lp in scaled_layer_probs.values():
+            lp[1, :] = 1 - (1 - lp[1, :]) ** dt  # Female participation
+            lp[2, :] = 1 - (1 - lp[2, :]) ** dt  # Male participation
+        scaled_f_cross = 1 - (1 - f_cross_layer) ** dt
+        scaled_m_cross = 1 - (1 - m_cross_layer) ** dt
+
         # Loop over layers
         lno = 0
         for lkey in self.layer_keys():
@@ -471,9 +481,9 @@ class People(hpb.BasePeople):
                 is_female=self.is_female,
                 is_active=self.is_active,
                 mixing=mixing[lkey],
-                layer_probs=layer_probs[lkey],
-                f_cross_layer=f_cross_layer,
-                m_cross_layer=m_cross_layer,
+                layer_probs=scaled_layer_probs[lkey],
+                f_cross_layer=scaled_f_cross,
+                m_cross_layer=scaled_m_cross,
                 durations=dur_pship[lkey],
                 acts=acts[lkey],
                 age_act_pars=age_act_pars[lkey],
@@ -942,31 +952,6 @@ class People(hpb.BasePeople):
         )  # These are not indices, so they scale differently
 
     # %% Methods to make events occur (death, infection, others TBC)
-    def make_naive(self, inds):
-        """
-        Make a set of people naive. This is used during dynamic resampling.
-
-        Args:
-            inds (array): list of people to make naive
-        """
-        for key in self.meta.states:
-            if key in ["susceptible"]:
-                self[key][:, inds] = True
-            elif key in ["other_dead"]:
-                self[key][inds] = False
-            else:
-                self[key][:, inds] = False
-
-        # Reset immunity
-        for key in self.meta.imm_states:
-            self[key][:, inds] = 0
-
-        # Reset dates
-        for key in self.meta.dates + self.meta.durs:
-            self[key][:, inds] = np.nan
-
-        return
-
     def infect(self, inds, g=None, layer=None):
         """
         Infect people and determine their eventual outcomes.
@@ -1049,6 +1034,14 @@ class People(hpb.BasePeople):
         f_inds = hpu.itruei(self.is_female, inds)
         m_inds = hpu.itruei(self.is_male, inds)
 
+        # Count new precin cases (females entering the precin state)
+        if layer != "seed_infection" and layer != "reactivation" and len(f_inds) > 0:
+            self.flows["precins"] += self.scale_flows(f_inds)
+            self.genotype_flows["precins"][g] += self.scale_flows(f_inds)
+            self.age_flows["precins"] += np.histogram(
+                self.age[f_inds], bins=self.age_bin_edges, weights=self.scale[f_inds]
+            )[0]
+
         # Compute disease progression for females
         if len(f_inds) > 0:
             gpars = self.pars["genotype_pars"][g]
@@ -1168,10 +1161,12 @@ class People(hpb.BasePeople):
 
             intro = f"\nThis is the story of {uid}, a {p.age:.0f} year old {sex}."
             intro += f"\n{uid} became sexually active at age {p.debut:.0f}."
-            if not p.susceptible:
-                if ~np.isnan(p.date_infectious):
+            if not np.all(p.susceptible):
+                dates_inf = p.date_infectious
+                if np.any(~np.isnan(dates_inf)):
+                    earliest = np.nanmin(dates_inf)
                     print(
-                        f"{intro}\n{uid} contracted HPV on timestep {p.date_infectious} of the simulation."
+                        f"{intro}\n{uid} contracted HPV on timestep {earliest:.0f} of the simulation."
                     )
                 else:
                     print(f"{intro}\n{uid} did not contract HPV during the simulation.")
@@ -1196,13 +1191,14 @@ class People(hpb.BasePeople):
             events = []
 
             dates = {
-                "date_HPV_clearance": "HPV cleared",
+                "date_clearance": "HPV cleared",
             }
 
             for attribute, message in dates.items():
                 date = getattr(p, attribute)
-                if not np.isnan(date):
-                    events.append((date, message))
+                if np.any(~np.isnan(date)):
+                    earliest = np.nanmin(date)
+                    events.append((earliest, message))
 
             if len(events):
                 for timestep, event in sorted(events, key=lambda x: x[0]):

{hpvsim-2.2.7 → hpvsim-2.3.0}/hpvsim/population.py

@@ -110,6 +110,17 @@ def make_people(sim, popdict=None, reset=False, verbose=None, use_age_data=True,
         # Create the contacts
         lkeys = sim['acts'].keys() # TODO: consider a more robust way to do this
         if microstructure in ['random', 'default']:
+            # Scale annual participation rates to per-timestep probabilities so
+            # that initial-network formation matches ongoing per-timestep rates
+            # (issue #13). Matches the scaling in People.create_partnerships.
+            dt = sim['dt']
+            scaled_layer_probs = {k: v.copy() for k, v in sim['layer_probs'].items()}
+            for lp in scaled_layer_probs.values():
+                lp[1, :] = 1 - (1 - lp[1, :]) ** dt
+                lp[2, :] = 1 - (1 - lp[2, :]) ** dt
+            scaled_f_cross = 1 - (1 - sim['f_cross_layer']) ** dt
+            scaled_m_cross = 1 - (1 - sim['m_cross_layer']) ** dt
+
             contacts = dict()
             current_partners = np.zeros((len(lkeys),n_agents))
             lno=0
@@ -117,7 +128,7 @@ def make_people(sim, popdict=None, reset=False, verbose=None, use_age_data=True,
                 contacts[lkey], current_partners,_,_ = make_contacts(
                     lno=lno, tind=0, partners=partners[lno,:], current_partners=current_partners, ages=ages,
                     debuts=debuts, is_female=is_female, is_active=is_active, mixing=sim['mixing'][lkey],
-                    layer_probs=sim['layer_probs'][lkey], f_cross_layer=sim['f_cross_layer'], m_cross_layer=sim['m_cross_layer'],
+                    layer_probs=scaled_layer_probs[lkey], f_cross_layer=scaled_f_cross, m_cross_layer=scaled_m_cross,
                     durations=sim['dur_pship'][lkey], acts=sim['acts'][lkey], age_act_pars=sim['age_act_pars'][lkey],
                     cluster=cluster, add_mixing=sim['add_mixing'], **kwargs
                 )

{hpvsim-2.2.7 → hpvsim-2.3.0}/hpvsim/sim.py

@@ -794,6 +794,8 @@ class Sim(hpb.BaseSim):
         for lkey, layer in people.contacts.items():
 
             sus = people.susceptible.copy() # for each layer, update who's still susceptible
+            if not hasattr(self, '_transmission_log'):
+                self._transmission_log = []
 
             # Shorten variables
             f = layer['f']
@@ -822,6 +824,8 @@ class Sim(hpb.BaseSim):
                     transmissions = (np.random.random(len(betas)) < betas).nonzero()[0] # Apply probabilities to determine partnerships in which transmission occurred
                     target_inds   = targets[transmissions] # Extract indices of those who got infected
                     target_inds, unique_inds = np.unique(target_inds, return_index=True)  # Due to multiple partnerships, some people will be counted twice; remove them
+                    source_inds = sources[transmissions[unique_inds]]
+                    self._transmission_log.append((source_inds, target_inds, lkey, g))
                     people.infect(inds=target_inds, g=g, layer=lkey)  # Infect people
 
         # Determine if there are any reactivated infections on this timestep

{hpvsim-2.2.7 → hpvsim-2.3.0}/hpvsim/version.py

@@ -4,6 +4,6 @@ Version and license information.
 
 __all__ = ['__version__', '__versiondate__', '__license__']
 
-__version__ = '2.2.7'
-__versiondate__ = '2026-04-22'
+__version__ = '2.3.0'
+__versiondate__ = '2026-04-20'
 __license__ = f'HPVsim {__version__} ({__versiondate__}) — © 2023-2026 by the Gates Foundation'

{hpvsim-2.2.7 → hpvsim-2.3.0/hpvsim.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: hpvsim
-Version: 2.2.7
+Version: 2.3.0
 Summary: HPVsim: Human Papillomavirus Simulator
 Author: Robyn Stuart, Jamie Cohen, Cliff Kerr, Romesh Abeysuriya, Mariah Boudreau, Daniel Klein, Hao Hu
 Maintainer-email: HPVsim Team <info@hpvsim.org>

{hpvsim-2.2.7 → hpvsim-2.3.0}/hpvsim.egg-info/SOURCES.txt

@@ -82,6 +82,7 @@ tests/baseline.json
 tests/benchmark.json
 tests/benchmark_profile.py
 tests/check_coverage
+tests/generate_v2_baselines.py
 tests/hpvsim_v1.2.2.yml
 tests/pytest.ini
 tests/requirements.txt
@@ -92,9 +93,15 @@ tests/test_baselines.py
 tests/test_calibration.py
 tests/test_data.py
 tests/test_event_schedule.py
+tests/test_immunity.py
 tests/test_interventions.py
+tests/test_misc.py
+tests/test_parameters.py
+tests/test_people.py
 tests/test_run.py
 tests/test_sim.py
+tests/test_utils.py
+tests/test_v2_regression.py
 tests/update_baseline
 tests/devtests/check_hiv_data.py
 tests/devtests/hpv_test_pars.csv

hpvsim-2.3.0/tests/baseline.json

@@ -0,0 +1,61 @@
+{
+  "summary": {
+    "infections": 25397265.0,
+    "precins": 8005294.0,
+    "cins": 1208087.46875,
+    "cancers": 21194.517578125,
+    "cancer_deaths": 21194.517578125,
+    "reinfections": 19518902.5,
+    "reactivations": 0.0,
+    "n_susceptible": 586788800.0,
+    "n_infectious": 32725578.0,
+    "n_inactive": 187010.4375,
+    "n_normal": 306995072.0,
+    "n_cin": 2937310.75,
+    "n_cancerous": 187010.4375,
+    "n_infected": 32912586.0,
+    "n_abnormal": 3124321.0,
+    "n_latent": 0.0,
+    "n_precin": 12149445.0,
+    "n_screened": 2515913.5,
+    "n_cin_treated": 0.0,
+    "n_cancer_treated": 0.0,
+    "n_vaccinated": 1284138.375,
+    "n_tx_vaccinated": 11888876.0,
+    "hpv_incidence": 0.021640891066768825,
+    "cin_incidence": 0.008250109718770863,
+    "cancer_incidence": 14.207040688510245,
+    "births": 7941710.1673,
+    "other_deaths": 3009621.5,
+    "migration": 299216.7096,
+    "asr_cancer_incidence": 18.621130304934443,
+    "asr_cancer_mortality": 19.734067713347496,
+    "new_vaccinated": 99738.90625,
+    "cum_vaccinated": 1371409.9609375,
+    "new_doses": 99738.90625,
+    "cum_doses": 1408812.05078125,
+    "new_txvx_doses": 12327728.0,
+    "new_tx_vaccinated": 11900097.25,
+    "cum_txvx_doses": 12327728.0,
+    "cum_tx_vaccinated": 11900097.25,
+    "new_screens": 2580744.5,
+    "new_screened": 2515914.125,
+    "new_cin_treatments": 0.0,
+    "new_cin_treated": 0.0,
+    "new_cancer_treatments": 0.0,
+    "new_cancer_treated": 0.0,
+    "cum_screens": 2580744.5,
+    "cum_screened": 2515914.125,
+    "cum_cin_treatments": 0.0,
+    "cum_cin_treated": 0.0,
+    "cum_cancer_treatments": 0.0,
+    "cum_cancer_treated": 0.0,
+    "cancer_mortality": 14.189253574665305,
+    "n_alive": 306995072.0,
+    "cdr": 0.00980348472824997,
+    "cbr": 0.025869178015013868,
+    "hpv_prevalence": 0.10659968509201347,
+    "precin_prevalence": 0.040668902998381044,
+    "cin_prevalence": 0.019664635869021486
+  }
+}

{hpvsim-2.2.7 → hpvsim-2.3.0}/tests/benchmark.json

@@ -1,7 +1,7 @@
 {
   "time": {
-    "initialize": 0.015,
-    "run": 1.809
+    "initialize": 0.011,
+    "run": 1.078
   },
   "parameters": {
     "n_agents": 10000,
@@ -10,5 +10,5 @@
     "n_interventions": 7,
     "n_analyzers": 0
   },
-  "cpu_performance": 0.3104645325856668
+  "cpu_performance": 0.5786863801131209
 }