slight reorganization, minor fixes of nothings, add tests

docs/src/canonical.jl

@@ -24,8 +24,8 @@
 
 # ## Defining the model
 #
-# For convenience in the later explanation, we list the whole definition of the
-# module here:
+# For convenience in the later explanation, we list the definition of the
+# interface for `AbstractFBCModels.CanonicalModel.Model` here:
 
 # ##LIST src/canonical.jl
 
@@ -147,3 +147,46 @@ mktempdir() do dir
     A.save(m, path)
     A.run_fbcmodel_file_tests(Model, path, name = "small model")
 end;
+
+# ### Changing identifiers in the model
+#
+# For convenience, function [`identifier_map`](@ref) can be used to mass-update
+# all identifiers and references in a whole canonical model. This is useful
+# e.g. when exporting model formats that require certain formatting of the
+# identifiers, such as SBML. We demonstrate this by changing all identifiers in
+# the above model to have a "type prefix" -- reaction identifiers start with
+# `R_`, metabolites start with `M_`, etc.
+
+import AbstractFBCModels.CanonicalModel: identifier_map
+
+m2 = identifier_map(
+    m,
+    reaction_map = id -> "R_" * id,
+    metabolite_map = id -> "M_" * id,
+    gene_map = id -> "G_" * id,
+    compartment_map = id -> "C_" * id,
+    coupling_map = id -> "X_" * id,
+);
+
+@test all(startswith("R_"), keys(m2.reactions)) #src
+@test all(startswith("M_"), keys(m2.metabolites)) #src
+@test all(startswith("G_"), keys(m2.genes)) #src
+@test all(startswith("C_"), keys(m2.compartments)) #src
+@test all(startswith("X_"), keys(m2.couplings)) #src
+
+mktempdir() do dir #src
+    path = joinpath(dir, "model.canonical-serialized-fbc") #src
+    A.save(m2, path) #src
+    res = A.run_fbcmodel_file_tests(Model, path, name = "reidentified model") #src
+end #src
+
+# We can see that the identifiers have changed:
+
+m2.reactions
+@test "R_and_back" in keys(m2.reactions) #src
+
+# ...together with the references:
+
+m2.reactions["R_forward"].stoichiometry
+
+@test "M_m1" in keys(m2.reactions["R_forward"].stoichiometry) #src

src/AbstractFBCModels.jl

@@ -34,6 +34,10 @@ include("utils.jl")
 include("accessors.jl")
 include("io.jl")
 include("testing.jl")
+
+module CanonicalModel
 include("canonical.jl")
+include("canonical_utils.jl")
+end
 
 end # module AbstractFBCModel

src/canonical.jl

@@ -1,6 +1,4 @@
 
-module CanonicalModel
-
 using DocStringExtensions
 
 import ..AbstractFBCModels as A
@@ -237,75 +235,3 @@ function Base.convert(::Type{Model}, x::A.AbstractFBCModel)
         ),
     )
 end
-
-"""
-$(TYPEDSIGNATURES)
-
-Rebuild the given [`Model`](@ref) with all identifiers changed accordingly to
-the given individual identifier-mapping functions in arguments. These take a
-single `String` identifier as an argument, and return a new `String`
-identifier. By default, the identifier maps are `identity`, i.e., no
-identifiers are changed.
-
-The identifier maps *must be injective and pure within `identifier_map`*, i.e.,
-they must not create identifier name conflicts, and the results returned for a
-given argument must be the same during the whole call of `identifier_map`.
-Errors stemming from use of non-injective and impure identifier maps are not
-handled.
-
-Internal data structures are copied only by reference wherever possible.
-"""
-function identifier_map(
-    m::Model;
-    reaction_map = identity,
-    metabolite_map = identity,
-    gene_map = identity,
-    coupling_map = identity,
-    compartment_map = identity,
-)
-    return Model(;
-        reactions = Dict(
-            reaction_map(k) => Reaction(;
-                name = v.name,
-                lower_bound = v.lower_bound,
-                upper_bound = v.upper_bound,
-                stoichiometry = Dict(
-                    metabolite_map(sk) => sv for (sk, sv) in v.stoichiometry
-                ),
-                objective_coefficient = v.objective_coefficient,
-                gene_association_dnf = [
-                    String[gene_map(g) for g in gs] for gs in a.gene_association_dnf
-                ],
-                annotations = v.annotations,
-                notes = v.notes,
-            ) for (k, v) in m.reactions
-        ),
-        metabolites = Dict(
-            metabolite_map(k) => Metabolite(;
-                name = v.name,
-                compartment = isnothing(v.compartment) ? nothing :
-                              compartment_map(v.compartment),
-                formula = v.formula,
-                charge = v.charge,
-                balance = v.balance,
-                annotations = v.annotations,
-                notes = v.notes,
-            ) for (k, v) in m.metabolites
-        ),
-        genes = Dict(gene_map(k) => v for (k, v) in m.genes),
-        couplings = Dict(
-            coupling_map(k) => Coupling(;
-                name = v.name,
-                reaction_weights = Dict(
-                    reaction_map(rk) => rv for (rk, rv) in v.reaction_weights
-                ),
-                lower_bound = v.lower_bound,
-                upper_bound = v.upper_bound,
-                annotations = v.annotations,
-                notes = v.notes,
-            ) for (k, v) in m.couplings
-        ),
-    )
-end
-
-end # module CanonicalModel

src/canonical_utils.jl

@@ -0,0 +1,71 @@
+
+"""
+$(TYPEDSIGNATURES)
+
+Rebuild the given [`Model`](@ref) with all identifiers changed accordingly to
+the given individual identifier-mapping functions in arguments. These take a
+single `String` identifier as an argument, and return a new `String`
+identifier. By default, the identifier maps are `identity`, i.e., no
+identifiers are changed.
+
+The identifier maps *must be injective and pure within `identifier_map`*, i.e.,
+they must not create identifier name conflicts, and the results returned for a
+given argument must be the same during the whole call of `identifier_map`.
+Errors stemming from use of non-injective and impure identifier maps are not
+handled.
+
+Internal data structures are copied only by reference wherever possible.
+"""
+function identifier_map(
+    m::Model;
+    reaction_map = identity,
+    metabolite_map = identity,
+    gene_map = identity,
+    coupling_map = identity,
+    compartment_map = identity,
+)
+    return Model(;
+        reactions = Dict(
+            reaction_map(k) => Reaction(;
+                name = v.name,
+                lower_bound = v.lower_bound,
+                upper_bound = v.upper_bound,
+                stoichiometry = Dict(
+                    metabolite_map(sk) => sv for (sk, sv) in v.stoichiometry
+                ),
+                objective_coefficient = v.objective_coefficient,
+                gene_association_dnf = isnothing(v.gene_association_dnf) ? nothing :
+                                       [
+                    String[gene_map(g) for g in gs] for gs in v.gene_association_dnf
+                ],
+                annotations = v.annotations,
+                notes = v.notes,
+            ) for (k, v) in m.reactions
+        ),
+        metabolites = Dict(
+            metabolite_map(k) => Metabolite(;
+                name = v.name,
+                compartment = isnothing(v.compartment) ? nothing :
+                              compartment_map(v.compartment),
+                formula = v.formula,
+                charge = v.charge,
+                balance = v.balance,
+                annotations = v.annotations,
+                notes = v.notes,
+            ) for (k, v) in m.metabolites
+        ),
+        genes = Dict(gene_map(k) => v for (k, v) in m.genes),
+        couplings = Dict(
+            coupling_map(k) => Coupling(;
+                name = v.name,
+                reaction_weights = Dict(
+                    reaction_map(rk) => rv for (rk, rv) in v.reaction_weights
+                ),
+                lower_bound = v.lower_bound,
+                upper_bound = v.upper_bound,
+                annotations = v.annotations,
+                notes = v.notes,
+            ) for (k, v) in m.couplings
+        ),
+    )
+end