[MRG] create ComputeParameters.from_manifest_row in support of `sou…

…rmash sketch fromfile` (#1886)

* add ComputeParameters.from_manifest_row

* moar

* add tests for ComputeParameters.from_manifest_row

src/sourmash/command_compute.py

@@ -328,6 +328,32 @@ def __init__(self, ksizes, seed, protein, dayhoff, hp, dna, num_hashes, track_ab
         self.track_abundance = track_abundance
         self.scaled = scaled
 
+    @classmethod
+    def from_manifest_row(cls, row):
+        "convert a CollectionManifest row into a ComputeParameters object"
+        is_dna = is_protein = is_dayhoff = is_hp = False
+        if row['moltype'] == 'DNA':
+            is_dna = True
+        elif row['moltype'] == 'protein':
+            is_protein = True
+        elif row['moltype'] == 'hp':
+            is_hp = True
+        elif row['moltype'] == 'dayhoff':
+            is_dayhoff = True
+        else:
+            assert 0
+
+        if is_dna:
+            ksize = row['ksize']
+        else:
+            ksize = row['ksize'] * 3
+
+        p = cls([ksize], 42, is_protein, is_dayhoff, is_hp, is_dna,
+                row['num'], row['with_abundance'], row['scaled'])
+
+        return p
+
+
     def to_param_str(self):
         "Convert object to equivalent params str."
         pi = []

src/sourmash/command_sketch.py

@@ -66,14 +66,8 @@ def _parse_params_str(params_str):
             if len(item) < 6 or item[4] != '=':
                 raise ValueError("seed takes a parameter, e.g. 'seed=42'")
             params['seed'] = int(item[5:])
-        elif item == 'protein':
-            moltype = 'protein'
-        elif item == 'dayhoff':
-            moltype = 'dayhoff'
-        elif item == 'hp':
-            moltype = 'hp'
-        elif item == 'dna':
-            moltype = 'dna'
+        elif item in ('protein', 'dayhoff', 'hp', 'dna'):
+            moltype = item
         else:
             raise ValueError(f"unknown component '{item}' in params string")
 

tests/test_sourmash_sketch.py

@@ -433,6 +433,77 @@ def test_compute_parameters_to_param_str(input_param_str, expected_output):
                                                   expected_output)
 
 
+def test_manifest_row_to_compute_parameters_1():
+    # test ComputeParameters.from_manifest_row with moltype 'DNA'
+    row = dict(moltype='DNA',
+               ksize=21,
+               num=0, scaled=1000,
+               with_abundance=1)
+    p = ComputeParameters.from_manifest_row(row)
+    assert p.dna
+    assert not p.protein
+    assert not p.dayhoff
+    assert not p.hp
+    assert p.num_hashes == 0
+    assert p.scaled == 1000
+    assert p.ksizes == [21]
+    assert p.track_abundance
+    assert p.seed == 42
+
+
+def test_manifest_row_to_compute_parameters_2():
+    # test ComputeParameters.from_manifest_row with moltype 'protein'
+    row = dict(moltype='protein',
+               ksize=10,
+               num=0, scaled=200,
+               with_abundance=1)
+    p = ComputeParameters.from_manifest_row(row)
+    assert not p.dna
+    assert p.protein
+    assert not p.dayhoff
+    assert not p.hp
+    assert p.num_hashes == 0
+    assert p.scaled == 200
+    assert p.ksizes == [30]
+    assert p.track_abundance
+    assert p.seed == 42
+
+
+def test_manifest_row_to_compute_parameters_3():
+    # test ComputeParameters.from_manifest_row with moltype 'dayhoff'
+    row = dict(moltype='dayhoff',
+               ksize=12,
+               num=0, scaled=200,
+               with_abundance=0)
+    p = ComputeParameters.from_manifest_row(row)
+    assert not p.dna
+    assert not p.protein
+    assert p.dayhoff
+    assert not p.hp
+    assert p.num_hashes == 0
+    assert p.scaled == 200
+    assert p.ksizes == [36]
+    assert not p.track_abundance
+    assert p.seed == 42
+
+
+def test_manifest_row_to_compute_parameters_4():
+    # test ComputeParameters.from_manifest_row with moltype 'hp'
+    row = dict(moltype='hp',
+               ksize=32,
+               num=0, scaled=200,
+               with_abundance=0)
+    p = ComputeParameters.from_manifest_row(row)
+    assert not p.dna
+    assert not p.protein
+    assert not p.dayhoff
+    assert p.hp
+    assert p.num_hashes == 0
+    assert p.scaled == 200
+    assert p.ksizes == [96]
+    assert not p.track_abundance
+    assert p.seed == 42
+
 ### command line tests