causarray 0.0.1__py3-none-any.whl → 0.0.2__py3-none-any.whl

causarray/DR_estimation.py

@@ -69,8 +69,8 @@ def cross_fitting(
     pi_hat : array
         Estimated propensity score.
     '''
-    func_ps, params_ps = _get_func_ps(ps_model, **kwargs)
-    params_glm = _filter_params(fit_glm, kwargs)
+    func_ps, params_ps = _get_func_ps(ps_model, verbose=verbose, **kwargs)
+    params_glm = _filter_params(fit_glm, {**kwargs, 'verbose': verbose})
 
     if verbose:
         pprint.pprint(params_ps)
@@ -97,6 +97,7 @@ def cross_fitting(
         Y_train, Y_test = Y[train_index], Y[test_index]
 
         if fit_pi:
+            if verbose: pprint.pprint('Fit propensity score models...')
             i_ctrl = (np.sum(A_train, axis=1) == 0.)
 
             pi = np.zeros_like(A_test, dtype=float)
@@ -111,13 +112,13 @@ def cross_fitting(
                 else:
                     pi[:,j] = func_ps(XA_train[i_cells], A_train[i_cells][:,j], XA_test)
 
+        if verbose: pprint.pprint('Fit outcome models...')
         # Fit GLM on training data and predict on test data
         res = fit_glm(Y_train, X_train, A_train, family=family, alpha=glm_alpha,
             impute=X_test, **params_glm)
         
         # Store results
-        if fit_pi:
-            pi_hat[test_index] = pi
+        if fit_pi: pi_hat[test_index] = pi
 
         Y_hat[test_index,:,:,0] = res[1][0]
         Y_hat[test_index,:,:,1] = res[1][1]

causarray/DR_learner.py

@@ -109,6 +109,7 @@ def compute_causal_estimand(
         Y_hat=Y_hat, pi_hat=pi_hat, random_state=random_state, verbose=verbose, **kwargs)
     pi_hat = pi_hat.reshape(*A.shape)
 
+    if verbose: pprint.pprint('Estimating AIPW mean...')
     # point estimation of the treatment effect
     _, etas = AIPW_mean(Y, np.stack([1-A, A], axis=-1), 
         Y_hat, np.stack([1-pi_hat, pi_hat], axis=-1), positive=True)

causarray/__about__.py

@@ -1 +1 @@
-__version__ = "0.0.1"
+__version__ = "0.0.2"

causarray/gcate_glm.py

@@ -49,13 +49,26 @@ def fit_glm(Y, X, A=None, family='gaussian', disp_family='poisson',
     impute : bool
         whether to impute potential outcomes and get predicted values
     offset : bool
-        whether to use log of sum of Y as offset    
+        whether to use log of sum of Y as offset
+
+    Returns
+    -------
+    B : array
+        d x p matrix of coefficients
+    Yhat : array
+        n x p x a matrix of predicted values
+    disp_glm : array
+        p x 1 vector of dispersion parameters
+    offsets : array
+        n x 1 vector of offsets
+    resid_deviance : array
+        n x p matrix of deviance residuals
     '''
     np.random.seed(random_state)
     
     if family not in ['gaussian', 'poisson', 'nb']:
         raise ValueError('Family not recognized')
-   
+
     d = X.shape[1]
 
     if A is None:
@@ -70,8 +83,6 @@ def fit_glm(Y, X, A=None, family='gaussian', disp_family='poisson',
             X_test = X
         X_test = np.c_[X,np.zeros_like(A)]
         X = np.c_[X,A]
-        # X_test = X.copy()
-
         a = A.shape[1]
 
     if offset is not None and offset is not False:
@@ -90,7 +101,7 @@ def fit_glm(Y, X, A=None, family='gaussian', disp_family='poisson',
     pprint.pprint('Fitting {} GLM{}...'.format(family, '' if offsets is None else ' with offset'))
     is_constant = np.all(X == X[0, :], axis=0)
     alpha[is_constant] = 0
-    # alpha[:-a] = 0
+
 
     families = {
         'gaussian': lambda disp: sm.families.Gaussian(),
@@ -122,11 +133,11 @@ def fit_glm(Y, X, A=None, family='gaussian', disp_family='poisson',
             Yhat_0 = np.zeros((Y.shape[0], a))
             Yhat_1 = np.zeros((Y.shape[0], a))
             if impute is not False:
-                for j in range(a):
+                for k in range(a):
                     X_test_copy = X_test.copy()
-                    Yhat_0[:,j] = mod.predict(X_test_copy, offset=offsets)                    
-                    X_test_copy[:, d+j] = 1  # Update the j-th column with all ones
-                    Yhat_1[:,j] = mod.predict(X_test_copy, offset=offsets)
+                    Yhat_0[:,k] = mod.predict(X_test_copy, offset=offsets)                    
+                    X_test_copy[:, d+k] = 1
+                    Yhat_1[:,k] = mod.predict(X_test_copy, offset=offsets)
             else:
                 Yhat_0[:,:] = Yhat_1[:,:] = mod.predict(X, offset=offsets).reshape(-1, a)
             
@@ -146,9 +157,11 @@ def fit_glm(Y, X, A=None, family='gaussian', disp_family='poisson',
 
     results = Parallel(n_jobs=n_jobs)(delayed(fit_model)(
         j, Y, X, offsets, family, disp_glm, impute, alpha) for j in tqdm(range(Y.shape[1])))
+    pprint.pprint('Fitting GLM done.')
+    if verbose: pprint.pprint('Fitting GLM done.')
 
     B, Yhat_0, Yhat_1, resid_deviance = zip(*results)
-    B = np.array(B)    
+    B = np.array(B)
     Yhat_0 = np.array(Yhat_0).transpose(1, 0, 2)
     Yhat_1 = np.array(Yhat_1).transpose(1, 0, 2)
     resid_deviance = np.array(resid_deviance).T
@@ -171,7 +184,6 @@ def estimate_disp(Y, X=None, A=None, Y_hat=None, disp_family='gaussian', offset=
     else:
         offsets = None
         sf = 1.
-    
 
     if Y_hat is None:        
         if verbose:
@@ -183,8 +195,7 @@ def estimate_disp(Y, X=None, A=None, Y_hat=None, disp_family='gaussian', offset=
         if disp_family=='gaussian':
             Y_norm = Y/sf
             reg = LinearRegression(fit_intercept=False).fit(X, Y_norm)
-            Y_hat = reg.predict(X)
-            # Y_hat = np.clip(Y_hat, 0, 1) * sf            
+            Y_hat = reg.predict(X)     
         elif disp_family=='poisson':
             Y_hat = fit_glm(Y, X, None, offset=offsets, family='poisson', impute=False, **kwargs)[1]      
             Y_hat /= sf

{causarray-0.0.1.dist-info → causarray-0.0.2.dist-info}/METADATA

@@ -1,7 +1,7 @@
 Metadata-Version: 2.4
 Name: causarray
-Version: 0.0.1
-Summary: causarray is a Python module for A Python package for simultaneous causal inference with an array of outcomes.
+Version: 0.0.2
+Summary: causarray is a Python module for simultaneous causal inference with an array of outcomes.
 Author-email: Jin-Hong Du <jinhongd@andrew.cmu.com>, Maya Shen <myshen@andrew.cmu.edu>, Hansruedi Mathys <mathysh@pitt.edu>, Kathryn Roeder <jinhongd@andrew.cmu.com>
 Maintainer-email: Jin-Hong Du <jinhongd@andrew.cmu.com>
 License: MIT License
@@ -23,25 +23,55 @@ Requires-Dist: statsmodels
 Requires-Dist: tqdm
 Description-Content-Type: text/markdown
 
+[![Documentation Status](https://readthedocs.org/projects/causarray/badge/?version=latest)](https://causarray.readthedocs.io/en/latest/?badge=latest)
+[![PyPI](https://img.shields.io/pypi/v/causarray?label=pypi)](https://pypi.org/project/causarray)
+[![PyPI-Downloads](https://img.shields.io/pepy/dt/causarray)](https://pepy.tech/project/causarray)
+
+
 # causarray
 
 Advances in single-cell sequencing and CRISPR technologies have enabled detailed case-control comparisons and experimental perturbations at single-cell resolution. However, uncovering causal relationships in observational genomic data remains challenging due to selection bias and inadequate adjustment for unmeasured confounders, particularly in heterogeneous datasets. To address these challenges, we introduce `causarray` [Du25], a doubly robust causal inference framework for analyzing array-based genomic data at both bulk-cell and single-cell levels. `causarray` integrates a generalized confounder adjustment method to account for unmeasured confounders and employs semiparametric inference with flexible machine learning techniques to ensure robust statistical estimation of treatment effects.
 
 
-# Requirements
+## Usage
 
-The dependencies for running `causarray` method are listed in `environment.yml` and can be installed by running
+We recommend using `causarray` in a conda environment:
+```cmd
+# create a new conda environment and install the necessary packages
+conda create -n causarray python=3.12 -y
+
+# activate the environment
+conda activate causarray
+```
 
+The module can be installed via PyPI:
 ```cmd
-PIP_NO_DEPS=1 conda env create -f environment.yml
+pip install causarray
 ```
 
+For `R` users, `reticulate` can be used to call `causarray` from `R`.
+The documentation and tutorials using both `Python` and `R` are available at [causarray.readthedocs.io](https://causarray.readthedocs.io/en/latest/).
+
+
 
+## Logs
+
+- [x] (2025-01-30) Python package released on PyPI
+- [x] (2025-02-01) code for reproducing figures in paper
+- [x] (2025-02-02) Tutorial for Python and R
+- [ ] Documentation
 
 
 <!-- 
 # Development
 
+The dependencies for running `causarray` method are listed in `environment.yml` and can be installed by running
+
+```cmd
+PIP_NO_DEPS=1 conda env create -f environment.yml
+```
+
+
 ## Build
 ```cmd
 git tag 0.0.0
@@ -59,12 +89,16 @@ python -m pytest tests/test_DR_learner.py
 
 ```cmd
 mkdir docs
-sphinx-quickstart
 cd docs
+sphinx-quickstart
+
 make html # sphinx-build source build
+
+
+rmarkdown::render("perturbseq.Rmd", rmarkdown::md_document(variant = "markdown_github"))
 ```
 -->
 
 
-# References
-[Du25] Jin-Hong Du, Maya Shen, Hansruedi Mathys, and Kathryn Roeder (2025). Causal differential expression analysis under unmeasured confounders with causarray. bioRxiv, 2025-01.
+## References
+[Du25] Jin-Hong Du, Maya Shen, Hansruedi Mathys, and Kathryn Roeder (2025). Causal differential expression analysis under unmeasured confounders with causarray. bioRxiv, 2025-01.

causarray-0.0.2.dist-info/RECORD

@@ -0,0 +1,14 @@
+causarray/DR_estimation.py,sha256=DqlrFEDj7RtAIUAqUf_6nXA2PRyk8qnouBjFzzIDxmo,9210
+causarray/DR_inference.py,sha256=0r5qxXTgvw-49ItP5TePC7KCDu4b8XrzgnlJfVWRcZo,5323
+causarray/DR_learner.py,sha256=4C1bAXr1L_yf8ZDqHMg40mkuQGqodSekqq6GjfqShAU,9461
+causarray/__about__.py,sha256=QvlVh4JTl3JL7jQAja76yKtT-IvF4631ASjWY1wS6AQ,22
+causarray/__init__.py,sha256=bTyCz9EY3NVccFb472w7gqMxrmidBsxHZjlsfMawPsk,657
+causarray/gcate.py,sha256=pnx8O-OnJl6VDjCF1RkgCuruQJVymTrGljc1V0jnRsc,8478
+causarray/gcate_glm.py,sha256=bqDEGjw9K089RA1ynrl_QbLgynKrl1VVMTKlf3wbS7Y,9213
+causarray/gcate_likelihood.py,sha256=srs2ql2uEhgn386w1lES41E_Gb0EKHT5YDIW-25YDec,4797
+causarray/gcate_opt.py,sha256=w3KWQiBfwek1O8uslFXPoDBol3XWj0U2yBYZ4WmMJxM,10589
+causarray/utils.py,sha256=fTrah_QX1C6FF6p2i-O1Y9_tljdjYcmxLkWE0tCbrJU,6765
+causarray-0.0.2.dist-info/METADATA,sha256=oBmoweLK43F9CaLag5c-c4GTWUSneRZRdY0tjoK2GaU,3547
+causarray-0.0.2.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+causarray-0.0.2.dist-info/licenses/LICENSE,sha256=4zFElAgYMtL4dKsu0A1p70cEo2SXRsGEWmJFdFb2hNg,1067
+causarray-0.0.2.dist-info/RECORD,,

causarray-0.0.1.dist-info/RECORD

@@ -1,14 +0,0 @@
-causarray/DR_estimation.py,sha256=wc8jsz1XjsXX5Vp6grbwvEYV8j1MHB-VA70wYbyYUEw,9050
-causarray/DR_inference.py,sha256=0r5qxXTgvw-49ItP5TePC7KCDu4b8XrzgnlJfVWRcZo,5323
-causarray/DR_learner.py,sha256=DdgGbVmvNbF-TulF5XGMvf5mf7Exd2uPPOoBeyRf4Ug,9404
-causarray/__about__.py,sha256=sXLh7g3KC4QCFxcZGBTpG2scR7hmmBsMjq6LqRptkRg,22
-causarray/__init__.py,sha256=bTyCz9EY3NVccFb472w7gqMxrmidBsxHZjlsfMawPsk,657
-causarray/gcate.py,sha256=pnx8O-OnJl6VDjCF1RkgCuruQJVymTrGljc1V0jnRsc,8478
-causarray/gcate_glm.py,sha256=lpVvsH2lDsmxvqzrXA9sbRf7bZed4of_rtM10eZHE2s,8956
-causarray/gcate_likelihood.py,sha256=srs2ql2uEhgn386w1lES41E_Gb0EKHT5YDIW-25YDec,4797
-causarray/gcate_opt.py,sha256=w3KWQiBfwek1O8uslFXPoDBol3XWj0U2yBYZ4WmMJxM,10589
-causarray/utils.py,sha256=fTrah_QX1C6FF6p2i-O1Y9_tljdjYcmxLkWE0tCbrJU,6765
-causarray-0.0.1.dist-info/METADATA,sha256=AoarxDPfvp0qQv1PSSkRwZsNGaEfcPTn0lj6DJ1SirY,2425
-causarray-0.0.1.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-causarray-0.0.1.dist-info/licenses/LICENSE,sha256=4zFElAgYMtL4dKsu0A1p70cEo2SXRsGEWmJFdFb2hNg,1067
-causarray-0.0.1.dist-info/RECORD,,