gotmap 0.1.0__tar.gz → 0.2.0__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: gotmap
3
- Version: 0.1.0
3
+ Version: 0.2.0
4
4
  Summary: Query-to-Reference Atlas Mapping for Single-Cell Data via Grassmann Optimal Transport
5
5
  Author: GOTMap Developers
6
6
  License: MIT
@@ -113,7 +113,7 @@ label_transfer_confidence(labelled, label_key="cell_type")
113
113
  ```python
114
114
  from gotmap import GOTMap
115
115
 
116
- mapper = GOTMap(ref_adata=ref, pca_key="X_pca", k=30, p=3, reg=0.1)
116
+ mapper = GOTMap(ref_adata=ref, use_rep="X_pca", k=30, p=3, reg=0.1)
117
117
  mapper.fit(query)
118
118
 
119
119
  # neural-network projection
@@ -96,7 +96,7 @@ label_transfer_confidence(labelled, label_key="cell_type")
96
96
  ```python
97
97
  from gotmap import GOTMap
98
98
 
99
- mapper = GOTMap(ref_adata=ref, pca_key="X_pca", k=30, p=3, reg=0.1)
99
+ mapper = GOTMap(ref_adata=ref, use_rep="X_pca", k=30, p=3, reg=0.1)
100
100
  mapper.fit(query)
101
101
 
102
102
  # neural-network projection
@@ -19,7 +19,7 @@ Note: torch is imported lazily (only when NN mapping is used).
19
19
  from .mapper import GOTMap
20
20
  from .update import FrozenReference
21
21
 
22
- __version__ = "0.1.0"
22
+ __version__ = "0.2.0"
23
23
 
24
24
  __all__ = [
25
25
  "GOTMap",
@@ -21,7 +21,7 @@ class GOTMap:
21
21
  ----------
22
22
  ref_adata : AnnData
23
23
  Reference atlas (well-annotated, stable).
24
- pca_key : str, default='X_pca'
24
+ use_rep : str, default='X_pca'
25
25
  Key in obsm for PCA coordinates.
26
26
  k : int, default=30
27
27
  Neighborhood size for Grassmann construction.
@@ -44,10 +44,10 @@ class GOTMap:
44
44
  Trained neural mapper (available after transform(method='nn')).
45
45
  """
46
46
 
47
- def __init__(self, ref_adata, pca_key="X_pca",
47
+ def __init__(self, ref_adata, use_rep="X_pca",
48
48
  k=30, p=3, reg=0.1, backend="batched", device=None):
49
49
  self.ref_adata = ref_adata.copy()
50
- self.pca_key = pca_key
50
+ self.use_rep = use_rep
51
51
  self.k = k
52
52
  self.p = p
53
53
  self.reg = reg
@@ -76,8 +76,8 @@ class GOTMap:
76
76
  -------
77
77
  self
78
78
  """
79
- H_X = self.ref_adata.obsm[self.pca_key].T.astype(np.float64)
80
- H_Y = query_adata.obsm[self.pca_key].T.astype(np.float64)
79
+ H_X = self.ref_adata.obsm[self.use_rep].T.astype(np.float64)
80
+ H_Y = query_adata.obsm[self.use_rep].T.astype(np.float64)
81
81
 
82
82
  got = GrassmannOT(
83
83
  k=self.k, p=self.p, reg=self.reg,
@@ -121,8 +121,8 @@ class GOTMap:
121
121
  def _transform_nn(self, query_adata, hidden_dim=128, epochs=200, lr=1e-3):
122
122
  from .nn_map import train_neural_mapper, apply_neural_mapper
123
123
 
124
- ref_pca = self.ref_adata.obsm[self.pca_key]
125
- query_pca = query_adata.obsm[self.pca_key]
124
+ ref_pca = self.ref_adata.obsm[self.use_rep]
125
+ query_pca = query_adata.obsm[self.use_rep]
126
126
 
127
127
  # anchor supervision
128
128
  anchor_ref = ref_pca[self.anchors_[:, 0]]
@@ -150,8 +150,8 @@ class GOTMap:
150
150
  def _transform_knn(self, query_adata, n_neighbors=15):
151
151
  from .knn_map import knn_projection
152
152
 
153
- ref_pca = self.ref_adata.obsm[self.pca_key]
154
- query_pca = query_adata.obsm[self.pca_key]
153
+ ref_pca = self.ref_adata.obsm[self.use_rep]
154
+ query_pca = query_adata.obsm[self.use_rep]
155
155
  projected = knn_projection(
156
156
  ref_pca, query_pca,
157
157
  self.anchors_, self.gamma_,
@@ -194,7 +194,7 @@ class GOTMap:
194
194
  ref_labels = self.ref_adata.obs[label_key].values
195
195
 
196
196
  pred_labels, conf = weighted_knn_labels(
197
- self.ref_adata.obsm[self.pca_key],
197
+ self.ref_adata.obsm[self.use_rep],
198
198
  query.obsm["X_gotmap"],
199
199
  ref_labels,
200
200
  gamma=self.gamma_,
@@ -233,7 +233,7 @@ class GOTMap:
233
233
  from .confidence import mapping_confidence
234
234
 
235
235
  return mapping_confidence(
236
- self.ref_adata.obsm[self.pca_key],
236
+ self.ref_adata.obsm[self.use_rep],
237
237
  query.obsm["X_gotmap"],
238
238
  self.anchors_,
239
239
  self.gamma_
@@ -15,7 +15,7 @@ from .mapper import GOTMap
15
15
  def map_query_to_reference(
16
16
  query,
17
17
  reference,
18
- pca_key="X_pca",
18
+ use_rep="X_pca",
19
19
  k=30,
20
20
  p=3,
21
21
  reg=0.1,
@@ -34,10 +34,10 @@ def map_query_to_reference(
34
34
  Parameters
35
35
  ----------
36
36
  query : AnnData
37
- Query data (must have .obsm[pca_key]).
37
+ Query data (must have .obsm[use_rep]).
38
38
  reference : AnnData
39
- Reference atlas (must have .obsm[pca_key] and .varm['PCs']).
40
- pca_key : str
39
+ Reference atlas (must have .obsm[use_rep] and .varm['PCs']).
40
+ use_rep : str
41
41
  Key for PCA coordinates.
42
42
  k, p, reg, backend : GrassmannOT parameters.
43
43
  method : str
@@ -62,7 +62,7 @@ def map_query_to_reference(
62
62
 
63
63
  mapper = GOTMap(
64
64
  ref_adata=reference,
65
- pca_key=pca_key,
65
+ use_rep=use_rep,
66
66
  k=k, p=p, reg=reg, backend=backend,
67
67
  )
68
68
  mapper.fit(query)
@@ -89,7 +89,7 @@ def transfer_labels(
89
89
  query,
90
90
  reference,
91
91
  label_key,
92
- pca_key="X_pca",
92
+ use_rep="X_pca",
93
93
  k=30,
94
94
  p=3,
95
95
  reg=0.1,
@@ -106,7 +106,7 @@ def transfer_labels(
106
106
  query, reference : AnnData
107
107
  label_key : str
108
108
  Column in reference.obs containing labels to transfer.
109
- pca_key : str
109
+ use_rep : str
110
110
  Key for PCA coordinates.
111
111
  method : str
112
112
  'weighted_knn' (default) or 'nn' (requires mapping first).
@@ -126,14 +126,14 @@ def transfer_labels(
126
126
 
127
127
  mapper = GOTMap(
128
128
  ref_adata=reference,
129
- pca_key=pca_key,
129
+ use_rep=use_rep,
130
130
  k=k, p=p, reg=reg, backend=backend,
131
131
  )
132
132
  mapped = mapper.fit(query).transform(query, method="knn")
133
133
 
134
134
  from .knn_map import weighted_knn_labels
135
135
  pred, conf = weighted_knn_labels(
136
- reference.obsm[pca_key],
136
+ reference.obsm[use_rep],
137
137
  mapped,
138
138
  reference.obs[label_key].values,
139
139
  gamma=mapper.gamma_,
@@ -152,7 +152,7 @@ def transfer_labels(
152
152
  def evaluate_mapping(
153
153
  query,
154
154
  reference,
155
- pca_key="X_pca",
155
+ use_rep="X_pca",
156
156
  k=30,
157
157
  p=3,
158
158
  reg=0.1,
@@ -167,7 +167,7 @@ def evaluate_mapping(
167
167
  Parameters
168
168
  ----------
169
169
  query, reference : AnnData
170
- pca_key : str
170
+ use_rep : str
171
171
  key_added : str
172
172
  Column name in query.obs for confidence scores.
173
173
 
@@ -178,7 +178,7 @@ def evaluate_mapping(
178
178
  """
179
179
  mapper = GOTMap(
180
180
  ref_adata=reference,
181
- pca_key=pca_key,
181
+ use_rep=use_rep,
182
182
  k=k, p=p, reg=reg, backend=backend,
183
183
  )
184
184
  quality = mapper.fit(query).evaluate_mapping(query)
@@ -43,16 +43,16 @@ class FrozenReference:
43
43
  ----------
44
44
  adata : AnnData
45
45
  Reference atlas (already preprocessed).
46
- pca_key : str
46
+ use_rep : str
47
47
  obsm key for PCA coordinates.
48
48
  use_highly_variable : bool
49
49
  Whether PCA was trained on highly variable genes.
50
50
  """
51
51
 
52
- def __init__(self, adata, pca_key="X_pca", use_highly_variable=True):
53
- if pca_key not in adata.obsm:
52
+ def __init__(self, adata, use_rep="X_pca", use_highly_variable=True):
53
+ if use_rep not in adata.obsm:
54
54
  raise ValueError(
55
- f"{pca_key} not found in adata.obsm. "
55
+ f"{use_rep} not found in adata.obsm. "
56
56
  f"Run sc.pp.pca(adata, n_comps=...) first."
57
57
  )
58
58
 
@@ -65,7 +65,7 @@ class FrozenReference:
65
65
  "This is required to project new data into the same PCA space."
66
66
  )
67
67
 
68
- self.pca_key = pca_key
68
+ self.use_rep = use_rep
69
69
  self.use_highly_variable = use_highly_variable
70
70
 
71
71
  self.PCs = adata.varm["PCs"].copy() # (n_genes_or_hvg, n_comps)
@@ -80,7 +80,7 @@ class FrozenReference:
80
80
  self.mean = X.mean(axis=0)
81
81
 
82
82
  # PCA coordinates of reference cells (frozen)
83
- self.pca = adata.obsm[pca_key].copy()
83
+ self.pca = adata.obsm[use_rep].copy()
84
84
 
85
85
  # gene names used for PCA (for alignment)
86
86
  self.var_names = adata.var_names.copy()
@@ -119,7 +119,7 @@ class FrozenReference:
119
119
  )
120
120
 
121
121
  meta = {
122
- "pca_key": self.pca_key,
122
+ "use_rep": self.use_rep,
123
123
  "use_highly_variable": self.use_highly_variable,
124
124
  "n_comps": int(self.n_comps),
125
125
  "n_obs": int(self.n_obs),
@@ -160,7 +160,7 @@ class FrozenReference:
160
160
 
161
161
  with open(os.path.join(path, "meta.json")) as f:
162
162
  meta = json.load(f)
163
- self.pca_key = meta["pca_key"]
163
+ self.use_rep = meta["use_rep"]
164
164
  self.use_highly_variable = meta["use_highly_variable"]
165
165
  self.version = meta["version"]
166
166
  self.history = meta.get("history", [])
@@ -309,7 +309,7 @@ class FrozenReference:
309
309
  sc.pp.pca(adata, n_comps=n_comps)
310
310
 
311
311
  self.PCs = adata.varm["PCs"].copy()
312
- self.pca = adata.obsm[self.pca_key].copy()
312
+ self.pca = adata.obsm[self.use_rep].copy()
313
313
  self.n_comps = n_comps
314
314
  self.n_obs = adata.n_obs
315
315
  self.var_names = adata.var_names.copy()
@@ -4,16 +4,16 @@ import numpy as np
4
4
  import scanpy as sc
5
5
 
6
6
 
7
- def check_pca(adata, pca_key="X_pca"):
8
- """Ensure that pca_key exists in adata.obsm."""
9
- if pca_key not in adata.obsm:
7
+ def check_pca(adata, use_rep="X_pca"):
8
+ """Ensure that use_rep exists in adata.obsm."""
9
+ if use_rep not in adata.obsm:
10
10
  raise ValueError(
11
- f"'{pca_key}' not found in adata.obsm. "
11
+ f"'{use_rep}' not found in adata.obsm. "
12
12
  f"Run sc.tl.pca(adata, n_comps=...) first."
13
13
  )
14
14
 
15
15
 
16
- def prep_adata(adata, n_comps=50, pca_key="X_pca"):
16
+ def prep_adata(adata, n_comps=50, use_rep="X_pca"):
17
17
  """
18
18
  Run standard preprocessing + PCA if not already done.
19
19
 
@@ -21,15 +21,15 @@ def prep_adata(adata, n_comps=50, pca_key="X_pca"):
21
21
  ----------
22
22
  adata : AnnData
23
23
  n_comps : int, default=50
24
- pca_key : str, default='X_pca'
24
+ use_rep : str, default='X_pca'
25
25
 
26
26
  Returns
27
27
  -------
28
28
  AnnData
29
- A copy with adata.obsm[pca_key] populated.
29
+ A copy with adata.obsm[use_rep] populated.
30
30
  """
31
31
  adata = adata.copy()
32
- if pca_key not in adata.obsm:
32
+ if use_rep not in adata.obsm:
33
33
  if "log1p" not in adata.uns:
34
34
  sc.pp.normalize_total(adata)
35
35
  sc.pp.log1p(adata)
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: gotmap
3
- Version: 0.1.0
3
+ Version: 0.2.0
4
4
  Summary: Query-to-Reference Atlas Mapping for Single-Cell Data via Grassmann Optimal Transport
5
5
  Author: GOTMap Developers
6
6
  License: MIT
@@ -113,7 +113,7 @@ label_transfer_confidence(labelled, label_key="cell_type")
113
113
  ```python
114
114
  from gotmap import GOTMap
115
115
 
116
- mapper = GOTMap(ref_adata=ref, pca_key="X_pca", k=30, p=3, reg=0.1)
116
+ mapper = GOTMap(ref_adata=ref, use_rep="X_pca", k=30, p=3, reg=0.1)
117
117
  mapper.fit(query)
118
118
 
119
119
  # neural-network projection
@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
4
4
 
5
5
  [project]
6
6
  name = "gotmap"
7
- version = "0.1.0"
7
+ version = "0.2.0"
8
8
  description = "Query-to-Reference Atlas Mapping for Single-Cell Data via Grassmann Optimal Transport"
9
9
  readme = "README.md"
10
10
  requires-python = ">=3.9"
File without changes
File without changes
File without changes
File without changes
File without changes
File without changes