Fix some typos

docs/userguide/campaigns.md

@@ -68,7 +68,7 @@ rec = campaign.recommend(batch_size=3)
 Calling the function returns a `DataFrame` with `batch_size` many rows, each
 representing a particular parameter configuration from the campaign's search space.
 Thus, the following might be a `DataFrame` returned by `recommend` in a search space
-with the three parameters `Categorial_1`, `Categorical_2` and `Num_disc_1`:
+with the three parameters `Categorical_1`, `Categorical_2` and `Num_disc_1`:
 
 |    | Categorical_1   | Categorical_2   |   Num_disc_1 |
 |---:|:----------------|:----------------|-------------:|

docs/userguide/serialization.md

@@ -252,7 +252,7 @@ BayBE objects typically appear as part of a larger object hierarchy.
 For instance, a
 {class}`SearchSpace <baybe.searchspace.core.SearchSpace>` can hold one or several
 {class}`Parameters <baybe.parameters.base.Parameter>`, just like an
-{class}`Objective <baybe.objectives.base.Objective>` can hold one ore several 
+{class}`Objective <baybe.objectives.base.Objective>` can hold one or several 
 {class}`Targets <baybe.targets.base.Target>`.
 This hierarchical structure can be directly replicated in the serialization string:
 

docs/userguide/surrogates.md

@@ -23,7 +23,7 @@ available within BayBE:
 In principle, the surrogate model does not need to be a persistent object during
 Bayesian optimization since each iteration performs a new fit anyway. However, for
 advanced study, such as investigating the posterior predictions, acquisition functions
-or feature importance, it can be useful to diretly extract the current surrogate model.
+or feature importance, it can be useful to directly extract the current surrogate model.
 
 For this, BayBE provides the ``get_surrogate`` method, which is available for the
 [``Campaign``](baybe.campaign.Campaign.get_surrogate) or for 

examples/Basics/recommenders.py

@@ -79,7 +79,7 @@
     "UCB",  # upper confidence bound with beta of 1.0
 ]
 
-# Note that the qvailability of the acquisition functions might depend on the `batch_size`:
+# Note that the availability of the acquisition functions might depend on the `batch_size`:
 #   - If `batch_size` is set to 1, all available acquisition functions can be chosen
 #   - If a larger value is chosen, only those that allow batching.
 #       That is, 'q'-variants of the acquisition functions must be chosen.

pyproject.toml

@@ -21,7 +21,7 @@ classifiers = [
 ]
 keywords = [
     "Active Learning",
-    "Bayesian Optmization",
+    "Bayesian Optimization",
     "Design of Experiments",
     "DOE",
     "Optimization",

tests/simulate_telemetry.py

@@ -97,7 +97,7 @@
     add_fake_measurements(dat, campaign.targets)
     campaign.add_measurements(dat)
 
-# Fake User1a - Adds recommenations before calling recommend
+# Fake User1a - Adds recommendations before calling recommend
 print("Fake User1a")
 os.environ[VARNAME_TELEMETRY_USERNAME] = "FAKE_USER_1a"
 campaign = Campaign(**config)

tests/test_iterations.py

@@ -90,7 +90,7 @@
 valid_nonmc_acqfs = [a() for a in get_subclasses(AcquisitionFunction) if not a.is_mc]
 
 # List of all hybrid recommenders with default attributes. Is extended with other lists
-# of hybird recommenders like naive ones or recommenders not using default arguments
+# of hybrid recommenders like naive ones or recommenders not using default arguments
 # TODO the TwoPhaseMetaRecommender below can be removed if the SeqGreedy recommender
 #  allows no training data
 valid_hybrid_recommenders = [