Remove deprecated types from manual page. (#848)

* Remove use of WrappedTreeSequence from manual page.

doc/long_vignettes/tskitconvert_vignette.md

@@ -23,6 +23,7 @@ This is intentional, as it keeps the execution times to a minimum while still il
 ---
 tags: [ "hide-input" ]
 ---
+import demes
 import fwdpy11
 ```
 
@@ -63,52 +64,6 @@ ts.dump("treefile.trees")
 We can also do a lot of analysis using this type.
 However, certain tasks would be easier if this type knew about the metadata generated by `fwdpy11`.
 
-### A lightweight wrapper
-
-The class {class}`fwdpy11.tskit_tools.WrappedTreeSequence` is a light wrapper around {class}`tskit.TreeSequence`.
-This class has properties to easily fetch the top-level metadata.
-We use these properties in the examples below.
-
-Instances of this wrapper class can be generated in several ways:
-
-* Directly:
-
-```{code-cell} python
-wrapper = fwdpy11.tskit_tools.WrappedTreeSequence(ts=ts)
-print(type(wrapper), type(wrapper.ts))
-```
-
-* Via a call to {func}`fwdpy11.tskit_tools.load`
-
-* By passing an option when exporting from `fwdpy11`:
-
-```{code-cell} python
-ts = pop.dump_tables_to_tskit(wrapped=True)
-print(type(ts), type(wrapper.ts))
-```
-
-To dump this wrapped type to a file:
-
-```python
-# not executed
-ts.ts.dump("treefile.trees")
-```
-
-In fact, all "`tskit` things" can be done by accessing {attr}`fwdpy11.tskit_tools.WrappedTreeSequence.ts`.
-
-Trying to create the wrapped type from a tree sequence not generated by `fwdpy11` is an error:
-
-```{code-cell} python
----
-tags: [raises-exception]
----
-import tskit
-tc = tskit.TableCollection(100.)
-ts = tc.tree_sequence()
-wrapper = fwdpy11.tskit_tools.WrappedTreeSequence(ts)
-```
-
-
 ## Metadata
 
 The `tskit` data model includes [metadata](https://tskit.dev/tskit/docs/stable/metadata.html).
@@ -153,20 +108,20 @@ For example, if you built the `rng` object and then simulated two replicates, th
 You do not know the state of the `rng` when the second simulation started.
 
 ```{code-cell} python
-ts = pop.dump_tables_to_tskit(wrapped=True)
-assert ts.seed is None
+ts = pop.dump_tables_to_tskit()
+assert "seed" not in ts.metadata
 ```
 
 ```{code-cell} python
-ts = pop.dump_tables_to_tskit(seed=54321, wrapped=True)
-assert ts.seed == 54321
+ts = pop.dump_tables_to_tskit(seed=54321)
+assert ts.metadata["seed"] == 54321
 ```
 
 ```{code-cell} python
 ---
 tags: [raises-exception]
 ---
-ts = pop.dump_tables_to_tskit(seed=-4616, wrapped=True)
+ts = pop.dump_tables_to_tskit(seed=-4616)
 ```
 
 ### Adding `ModelParams` objects
@@ -176,8 +131,8 @@ Including this is optional for several reasons, but a big one is that this class
 The examples listed below will work with types provided by `fwdpy11` (although we haven't tested them all!), but user-defined types may be more challenging.
 
 ```{code-cell} python
-ts = pop.dump_tables_to_tskit(wrapped=True)
-assert ts.model_params is None
+ts = pop.dump_tables_to_tskit()
+assert "model_params" not in ts.metadata
 ```
 
 Here is an example of Gaussian stabilizing selection around an optimum trait value of `0.0`:
@@ -199,8 +154,8 @@ params = fwdpy11.ModelParams(**pdict)
 We can pass the `params` object on when exporting the data to `tskit`:
 
 ```{code-cell} python
-ts = pop.dump_tables_to_tskit(model_params=params, wrapped=True)
-assert ts.model_params == params
+ts = pop.dump_tables_to_tskit(model_params=params)
+assert fwdpy11.ModelParams(**eval(ts.metadata["model_params"])) == params
 ```
 
 Some simulation designs make use of multiple `ModelParams` objects.
@@ -228,13 +183,13 @@ params2 = fwdpy11.ModelParams(**pdict2)
 We can pass both objects to `tskit` via a {class}`dict`:
 
 ```{code-cell} python
-ts = pop.dump_tables_to_tskit(model_params={'burnin': params, 'adaptation': params2}, wrapped=True)
-type(ts.model_params)
+ts = pop.dump_tables_to_tskit(model_params={'burnin': params, 'adaptation': params2})
+type(ts.metadata["model_params"])
 ```
 
 ```{code-cell} python
-assert ts.model_params['burnin'] == params
-assert ts.model_params['adaptation'] == params2
+assert fwdpy11.ModelParams(**eval(ts.metadata["model_params"]["burnin"])) == params
+assert fwdpy11.ModelParams(**eval(ts.metadata["model_params"]["adaptation"])) == params2
 ```
 
 ### Including a `demes` graph
@@ -301,18 +256,18 @@ migrations:
 
 ```{code-cell} python
 graph = gutenkunst()
-ts = pop.dump_tables_to_tskit(demes_graph=graph, wrapped=True)
+ts = pop.dump_tables_to_tskit(demes_graph=graph)
 ```
 
 ```{code-cell} python
-assert ts.demes_graph == graph
+assert demes.Graph.fromdict(ts.metadata["demes_graph"]) == graph
 ```
 
 Since this is an optional metadata field, accessing it will return `None` if no graph was provided:
 
 ```{code-cell} python
-ts = pop.dump_tables_to_tskit(wrapped=True)
-assert ts.demes_graph is None
+ts = pop.dump_tables_to_tskit()
+assert "demes_graph" not in ts.metadata
 ```
 
 ### User-defined metadata
@@ -323,8 +278,8 @@ For example, this could be a {class}`dict` containing information obtained from
 The simplest situation is to use a `dict` containing simple objects:
 
 ```{code-cell} python
-ts = pop.dump_tables_to_tskit(data={'x': 3}, wrapped=True)
-ts.data
+ts = pop.dump_tables_to_tskit(data={'x': 3})
+ts.metadata["data"]
 ```
 
 Any use of user-defined types will require conversion to string, however:
@@ -338,13 +293,13 @@ class MyData(object):
         return f"MyData(x={self.x})"
 
 # Send in the str representation:
-ts = pop.dump_tables_to_tskit(data=str(MyData(x=77)), wrapped=True)
+ts = pop.dump_tables_to_tskit(data=str(MyData(x=77)))
 ```
 
 Getting the data back out requires an `eval`:
 
 ```{code-cell} python
-assert eval(ts.data).x == 77
+assert eval(ts.metadata["data"]).x == 77
 ```
 
 :::{warning}
@@ -384,8 +339,8 @@ To actually illustrate the use of this metadata, we need to make sure that our `
 ```{code-cell} python
 # initialize a population with the right number of demes...
 multideme_pop = fwdpy11.DiploidPopulation([100]*len(pop_md), 1.)
-ts = multideme_pop.dump_tables_to_tskit(population_metadata=pop_md, wrapped=True)
-for pop in ts.ts.populations():
+ts = multideme_pop.dump_tables_to_tskit(population_metadata=pop_md)
+for pop in ts.populations():
     print(pop.metadata)
 ```
 

doc/misc/changelog.md

@@ -21,6 +21,7 @@ Dependencies
 
 * Deprecate `fwdpy11.tskit_tools.WrappedTreeSequence`
   PR {pr}`841`
+  PR {pr}`848`
 
 Back end changes: