Merge 1.0.0a2

README.md

@@ -1,22 +1,19 @@
-# LIANA+: an all-in-one framework for cell-cell communication <img src="https://raw.githubusercontent.com/saezlab/liana-py/dev/docs/source/_static/logo.png?raw=true" align="right" height="125">
+# LIANA+: an all-in-one cell-cell communication framework <img src="https://raw.githubusercontent.com/saezlab/liana-py/dev/docs/source/_static/logo.png?raw=true" align="right" height="125">
 
 <!-- badges: start -->
 [![main](https://github.com/saezlab/liana-py/actions/workflows/main.yml/badge.svg)](https://github.com/saezlab/liana-py/actions)
 [![GitHub issues](https://img.shields.io/github/issues/saezlab/liana-py.svg)](https://github.com/saezlab/liana-py/issues/)
 [![Documentation Status](https://readthedocs.org/projects/liana-py/badge/?version=latest)](https://liana-py.readthedocs.io/en/latest/?badge=latest)
 [![codecov](https://codecov.io/gh/saezlab/liana-py/branch/main/graph/badge.svg?token=TM0P29KKN5)](https://codecov.io/gh/saezlab/liana-py)
-[![Downloads](https://pepy.tech/badge/liana)](https://pepy.tech/project/liana)
+[![Downloads](https://static.pepy.tech/badge/liana)](https://pepy.tech/project/liana)
 <!-- badges: end -->
 
-LIANA+ is an efficient framework that integrates and extends existing methods and knowledge to study cell-cell communication in single-cell, spatially-resolved, and multi-modal omics data. It works in conjunction with the [scverse ecosystem](https://github.com/scverse), and it relies on [AnnData](https://github.com/scverse/anndata) & [MuData](https://github.com/scverse/mudata) objects as input.
+LIANA+ is a scalable framework that integrates and extends existing methods and knowledge to study cell-cell communication in single-cell, spatially-resolved, and multi-modal omics data. It works in conjunction with the [scverse ecosystem](https://github.com/scverse), and it relies on [AnnData](https://github.com/scverse/anndata) & [MuData](https://github.com/scverse/mudata) objects as input.
 
 <img src="https://raw.githubusercontent.com/saezlab/liana-py/main/docs/source/_static/abstract.png" width="700" align="center">
 
-
 ## Development & Contributions
 
-LIANA+ is still under development. Nevertheless, the API visible to the user should be stable, specifically `liana.method.sc`.
-
 We welcome suggestions, ideas, and contributions! Please use do not hesitate to contact us, or use the issues or the [LIANA+ Development project](https://github.com/orgs/saezlab/projects/16) to make suggestions.
 
 ## Tutorials
@@ -37,9 +34,9 @@ We welcome suggestions, ideas, and contributions! Please use do not hesitate to
 
 - [Multicellular programmes with MOFA](https://liana-py.readthedocs.io/en/latest/notebooks/mofacellular.html). Using MOFA to obtain coordinates gene expression programmes across samples and conditions, as done in [Ramirez et al., 2023](https://europepmc.org/article/ppr/ppr620471).
 
-- [LIANA with MOFA](https://liana-py.readthedocs.io/en/latest/notebooks/mofatalk.html). Using MOFA to infer intercellular communication programmes across samples and conditions, as initially proposed by cell2cell-Tensor.
+- [LIANA with MOFA](https://liana-py.readthedocs.io/en/latest/notebooks/mofatalk.html). Using MOFA to infer intercellular communication programmes across samples and conditions, as initially proposed by Tensor-cell2cell.
 
-- [LIANA with cell2cell-Tensor](https://liana-py.readthedocs.io/en/latest/notebooks/liana_c2c.html) to extract intercellular communication programmes across samples and conditions. Extensive tutorials combining LIANA & [cell2cell-Tensor](https://www.nature.com/articles/s41467-022-31369-2) are available [here](https://ccc-protocols.readthedocs.io/en/latest/index.html).
+- [LIANA with Tensor-cell2cell](https://liana-py.readthedocs.io/en/latest/notebooks/liana_c2c.html) to extract intercellular communication programmes across samples and conditions. Extensive tutorials combining LIANA & [Tensor-cell2cell](https://www.nature.com/articles/s41467-022-31369-2) are available [here](https://ccc-protocols.readthedocs.io/en/latest/index.html).
 
 
 ### Others

docs/source/index.rst

@@ -42,11 +42,6 @@ spatially-resolved, and multi-modal omics data.
 
    notebooks/basic_usage 
 
-.. toctree::
-   :maxdepth: 1
-   :hidden:
-   :caption: Multi-Sample Vignettes
-
    notebooks/liana_c2c
 
    notebooks/mofacellular

docs/source/release_notes.rst

@@ -1,7 +1,32 @@
 Release notes
 =============
 
-1.0.0a1 (30.07.2023)
+1.0.0a2 (19.09.2023)
+
+- Interactions names in `tileplot` and `dotplot` will now be sorted according to `orderby` when used; related to #55
+
+- Added `filter_view_markers` function to filter view markers considered background in MOFAcellular tutorial
+
+- Added `keep_stats` parameter to `adata_to_views` to enable pseudobulk stats to be kept.
+
+- Replace `intra_groupby` and `extra_groupby` with `maskby` in misty. 
+  The spots will now only be filtered according to `maskby`, such that both intra and extra both contain the same spots.
+  The extra views are multiplied by the spatial connectivities prior to masking and the model being fit
+
+- Merge MOFAcell improvements; related to #42 and #29
+
+- Targets with zero variance will no longer be modeled by misty.
+
+- Resolve #46 - refactored misty's pipeline
+
+- Resolved logging and package import verbosity issues related to #43
+
+- Iternal .obs['label'] placeholder renamed to the less generic .obs['@label']; related to #53
+
+- Minor Readme & tutorial text improvements.
+
+
+1.0.0a1 Biorxiv (30.07.2023)
 ---------------------------------------------------------
 
 - `positive_only` in bivariate metrics was renamed to `mask_negatives` will now mask only negative-negative/low-low interactions, and not negative-positive interactions.

liana/__init__.py

@@ -1,7 +1,7 @@
 import importlib.metadata
 __version__ = importlib.metadata.version("liana")
 
-from liana import method as mt, plotting as pl, resource as rs, multi as mu, utils as ut, testing
+from liana import _logging, method as mt, plotting as pl, resource as rs, multi as mu, utils as ut, testing
 
 # done after everything has been imported (adapted from scanpy)
 import sys

liana/_logging.py

@@ -0,0 +1,29 @@
+import logging
+from types import ModuleType
+
+logging.basicConfig(level=logging.INFO, format='%(message)s')
+
+def _logg(message, level='info', verbose=False):
+    """
+    Log a message with a specified logging level.
+
+    Args:
+        message (str): The message to log.
+        level (int or None): The logging level for the message (default is None).
+            If None or False, no logging is performed.
+    """
+    if verbose:
+        if level == "warn":
+            logging.warning(message)
+        elif level == "info":
+            logging.info(message)
+
+def _check_if_installed(package_name: str, custom_error_message: str = None) -> ModuleType:
+    try:
+        imported_module = __import__(package_name)
+        return imported_module
+    except ImportError:
+        if custom_error_message:
+            raise ImportError(custom_error_message)
+        else:
+            raise ImportError(f'{package_name} is not installed. Please install it with: pip install {package_name}')

liana/method/_pipe_utils/_get_mean_perms.py

@@ -42,12 +42,12 @@ def _get_means_perms(adata: anndata.AnnData,
         adata.X /= norm_factor
 
     # define labels and masks
-    labels = adata.obs.label.cat.categories
+    labels = adata.obs['@label'].cat.categories
     labels_mask = np.zeros((adata.shape[0], labels.shape[0]), dtype=bool)
 
     # populate masks shape(genes, labels)
     for ct_idx, label in enumerate(labels):
-        labels_mask[:, ct_idx] = adata.obs.label == label
+        labels_mask[:, ct_idx] = adata.obs['@label'] == label
 
     # Perm should be a cube /w dims: n_perms x idents x n_genes
     perms = _generate_perms_cube(adata.X, n_perms, labels_mask, seed, agg_fun, verbose)
@@ -79,7 +79,7 @@ def _generate_perms_cube(X, n_perms, labels_mask, seed, agg_fun, verbose):
 
 
 def _get_positions(adata, lr_res):
-    labels = adata.obs['label'].cat.categories
+    labels = adata.obs['@label'].cat.categories
 
     # get positions of each entity in the matrix
     ligand_pos = {entity: np.where(adata.var_names == entity)[0][0] for entity

liana/method/_pipe_utils/_pre.py

@@ -10,7 +10,7 @@
 from pandas import DataFrame, Index
 import scanpy as sc
 from scipy.sparse import csr_matrix, isspmatrix_csr
-
+from liana._logging import _logg
 
 def assert_covered(
         subset,
@@ -56,8 +56,7 @@ def assert_covered(
         )
         raise ValueError(msg + f" [{x_missing}] missing from {superset_name}")
 
-    if verbose & (prop_missing > 0):
-        print(f"{prop_missing:.2f} of entities in the resource are missing from the data.")
+    _logg(f"{prop_missing:.2f} of entities in the resource are missing from the data.", verbose=verbose & (prop_missing > 0))
 
 
 def prep_check_adata(adata: AnnData,
@@ -114,33 +113,29 @@ def prep_check_adata(adata: AnnData,
     msk_features = np.sum(adata.X, axis=0).A1 == 0
     n_empty_features = np.sum(msk_features)
     if n_empty_features > 0:
-        if verbose:
-            print(f"{n_empty_features} features of mat are empty, they will be removed.")
+        _logg(f"{n_empty_features} features of mat are empty, they will be removed.", level='warn', verbose=verbose)
         adata = adata[:, ~msk_features]
 
     # Check for empty samples
     msk_samples = adata.X.sum(axis=1).A1 == 0
     n_empty_samples = np.sum(msk_samples)
     if n_empty_samples > 0:
-        if verbose:
-            print(f"{n_empty_samples} samples of mat are empty, please remove them.")
+        _logg(f"{n_empty_samples} samples of mat are empty, they will be removed.", level='warn', verbose=verbose)
 
     # Check if log-norm
     _sum = np.sum(adata.X.data[0:100])
     if _sum == np.floor(_sum):
-        if verbose:
-            print("Make sure that normalized counts are passed!")
+        _logg("Make sure that normalized counts are passed!", level='warn', verbose=verbose)
 
     # Check for non-finite values
     if np.any(~np.isfinite(adata.X.data)):
-        raise ValueError(
-            "mat contains non finite values (nan or inf), please set them to 0 or remove them.")
+        raise ValueError("mat contains non finite values (nan or inf), please set them to 0 or remove them.")
 
     # Define idents col name
     if groupby is not None:
         if groupby not in adata.obs.columns:
             raise AssertionError(f"`{groupby}` not found in `adata.obs.columns`.")
-        adata.obs.loc[:, 'label'] = adata.obs[groupby]
+        adata.obs.loc[:, '@label'] = adata.obs[groupby]
 
         # Remove any cell types below X number of cells per cell type
         count_cells = adata.obs.groupby(groupby)[groupby].size().reset_index(name='count').copy()
@@ -151,9 +146,9 @@ def prep_check_adata(adata: AnnData,
             # remove lowly abundant identities
             msk = ~np.isin(adata.obs[[groupby]], lowly_abundant_idents)
             adata = adata[msk]
-            if verbose:
-                print("The following cell identities were excluded: {0}".format(
-                    ", ".join(lowly_abundant_idents)))
+            _logg("The following cell identities were excluded: {0}".format(", ".join(lowly_abundant_idents)),
+                 level='warn', verbose=verbose)
+
 
     check_vars(adata.var_names,
                complex_sep=complex_sep,
@@ -178,8 +173,8 @@ def check_vars(var_names, complex_sep, verbose=False) -> list:
     else:
         pass
 
-    if verbose & (len(var_issues) > 0):
-        print(f"Warning: {var_issues} contain `{complex_sep}`. Consider replacing those!")
+    _logg(f"{var_issues} contain `{complex_sep}`. Consider replacing those!",
+         verbose=verbose & (len(var_issues) > 0), level='warn')
 
 
 
@@ -243,24 +238,20 @@ def _choose_mtx_rep(adata, use_raw=False, layer=None, verbose=False) -> csr_matr
     if is_layer & use_raw:
         raise ValueError("Cannot specify `layer` and have `use_raw=True`.")
     if is_layer:
-        if verbose:
-            print(f"Using the `{layer}` layer!")
+        _logg(f"Using the `{layer}` layer!", verbose=verbose)
         X = adata.layers[layer]
     elif use_raw:
         if adata.raw is None:
             raise ValueError("`.raw` is not initialized!")
-        if verbose:
-            print("Using `.raw`!")
+        _logg("Using `.raw`!", verbose=verbose)
         X = adata.raw.X
     else:
-        if verbose:
-            print("Using `.X`!")
+        _logg("Using `.X`!", verbose=verbose)
         X = adata.X
 
     # convert to sparse csr matrix
     if not isspmatrix_csr(X):
-        if verbose:
-            print("Converting mat to CSR format")
+        _logg("Converting to sparse csr matrix!", verbose=verbose)
         X = csr_matrix(X)
 
     return X

liana/method/_pipe_utils/_reassemble_complexes.py

@@ -4,8 +4,7 @@
 from __future__ import annotations
 
 import pandas as pd
-import numpy as np
-
+from liana._logging import _logg
 
 def filter_reassemble_complexes(lr_res,
                                 _key_cols,
@@ -73,9 +72,10 @@ def filter_reassemble_complexes(lr_res,
     if duplicate_mask.any():
         # check if there are any non-equal subunit values
         if not lr_res[duplicate_mask].groupby(_key_cols)[complex_cols].transform(lambda x: x.duplicated(keep=False)).all().all():
-            print('Warning: there were duplicated subunits in the complexes. ' + 
-                  'The subunits were reduced to only the minimum expression subunit. ' +
-                  'However, there were subunits that were not the same within a complex. ')
+            _logg('There were duplicated subunits in the complexes. ' + 
+                 'The subunits were reduced to only the minimum expression subunit. ' +
+                 'However, there were subunits that were not the same within a complex. ',
+                 level='warn')
         lr_res = lr_res.drop_duplicates(subset=_key_cols, keep='first')
 
     return lr_res