U 3d @shdZddlZddlmZddlZddlZddlZddlZddl Z ddl m Z m Z m Z ddl mZddlZddlZddlmZddlmZddlZddlmZddlmZddlZddlZddlZdd lmZzeWnek rdZYnXdd lmZ dd lm!Z!dd lm"Z"dd lm#Z#ddlm$Z$ddlm%Z%ddl&Z'ddl(Z(ddl)Z)ddl*m+Z+m,Z,m-Z-ddl.m/Z/ddl0m1Z1m2Z2m3Z3ddl4m5Z5ddddddddddg Z6edZ7e7j8Z9ej:j;Z;e7jZ?e?Z@dd ZAdeBfd!d"ZCGd#d$d$ZDd%d&ZEd]d*dZd^d-d.ZFd_d/d0ZGzddlHZHeHjIjJe,d1d2ZKeHjIjJejLMd3d4kd5d2ZNeHjIjOe+d6d2ZPeHjIjOe-d7d2ZQeHjIjJe(jRjS d8d2ZTeHjIjJejUd9kd:d2ZVWneWk rYnXd;d<ZXd`d>d?ZYGd@dAdAZZdBdCZ[dDdEZ\dadGdHZ]dbdIdJZ^dKdLZ_dcdMdNZ`dddPdZadedQdRZbdfdSdTZcGdUdVdVejdZeGdWdXdXZfGdYdZdZZgGd[d\d\ZhdS)gzTesting utilities.N) check_outputSTDOUTCalledProcessError)TimeoutExpired)Iterable)Sequencewraps) signature)TestCase)assert_allclose)assert_almost_equal)assert_approx_equal)assert_array_equal)assert_array_almost_equal)assert_array_less)IS_PYPY _IS_32BIT#_in_unstable_openblas_configuration)check_classification_targets) check_arraycheck_is_fitted check_X_y)threadpool_info assert_raisesassert_raises_regexprr rrrr assert_run_python_scriptSkipTest__init__c Osvtjdd`}td|||}ttdr:dd|D}t|dkrhtd|jd d d |DfW5QRX|S) z? Parameters ---------- func *args **kw Trecordalways FutureWarningcSsg|]}|jtjk r|qS)categorynpZVisibleDeprecationWarning.0er#r#:/tmp/pip-unpacked-wheel-zrfo1fqw/sklearn/utils/_testing.py ks z&assert_no_warnings..rz"Got warnings when calling %s: [%s], css|]}t|VqdSN)str)r'warningr#r#r) psz%assert_no_warnings..) warningscatch_warnings simplefilterhasattrr%lenAssertionError__name__join)funcargskwwresultr#r#r)assert_no_warnings\s     r=cCsPt|tr,t|tr,|j}tdj|dn t|rBt|d|St|dSdS)a Context manager and decorator to ignore warnings. Note: Using this (in both variants) will clear all warnings from all python modules loaded. In case you need to test cross-module-warning-logging, this is not your tool of choice. Parameters ---------- obj : callable, default=None callable where you want to ignore the warnings. category : warning class, default=Warning The category to filter. If Warning, all categories will be muted. Examples -------- >>> import warnings >>> from sklearn.utils._testing import ignore_warnings >>> with ignore_warnings(): ... warnings.warn('buhuhuhu') >>> def nasty_warn(): ... warnings.warn('buhuhuhu') ... print(42) >>> ignore_warnings(nasty_warn)() 42 z'obj' should be a callable where you want to ignore warnings. You passed a warning class instead: 'obj={warning_name}'. If you want to pass a warning class to ignore_warnings, you should use 'category={warning_name}') warning_name)r$N) isinstancetype issubclassWarningr6 ValueErrorformatcallable_IgnoreWarnings)objr$r>r#r#r)ignore_warningsusrHc@s8eZdZdZddZddZddZdd Zd d Zd S) rFatImproved and simplified Python warnings context manager and decorator. This class allows the user to ignore the warnings raised by a function. Copied from Python 2.7.5 and modified as required. Parameters ---------- category : tuple of warning class, default=Warning The category to filter. By default, all the categories will be muted. cCs(d|_tjd|_d|_g|_||_dS)NTr0F)_recordsysmodules_module_enteredlogr$)selfr$r#r#r)rs  z_IgnoreWarnings.__init__cstfdd}|S)z.wrapperr)rOrSrTr#rRr)__call__sz_IgnoreWarnings.__call__cCsPg}|jr|d|jtjdk r4|d|jt|j}d|d|fS)Nz record=Truer0z module=%rz%s(%s)r+)rIappendrLrJrKr@r6r7)rOr9namer#r#r)__repr__s  z_IgnoreWarnings.__repr__cCsP|jrtd|d|_|jj|_|jdd|j_|jj|_td|j dS)NzCannot enter %r twiceTrP) rM RuntimeErrorrLfilters_filters showwarning _showwarningr0r2r$)rOr#r#r) __enter__s   z_IgnoreWarnings.__enter__cGs8|jstd||j|j_|j|j_g|jdd<dS)Nz%Cannot exit %r without entering first)rMrYr[rLrZr]r\rN)rOexc_infor#r#r)__exit__s    z_IgnoreWarnings.__exit__N) r6 __module__ __qualname____doc__rrUrXr^r`r#r#r#r)rFs     rFc Osz|||WnB|k rP}z$t|}||kr@td||fW5d}~XYn:Xt|trrddd|D}n|j}td||jfdS)aHelper function to test the message raised in an exception. Given an exception, a callable to raise the exception, and a message string, tests that the correct exception is raised and that the message is a substring of the error thrown. Used to test that the specific message thrown during an exception is correct. Parameters ---------- exceptions : exception or tuple of exception An Exception object. message : str The error message or a substring of the error message. function : callable Callable object to raise error. *args : the positional arguments to `function`. **kwargs : the keyword arguments to `function`. zRError message does not include the expected string: %r. Observed error message: %rNz or css|] }|jVqdSr,)r6r&r#r#r)r/sz'assert_raise_message..z%s not raised by %s)r-r5r?tupler7r6) exceptionsmessagefunctionr9rQr( error_messagenamesr#r#r)assert_raise_messages rjTc Cs^g}t|t|}}|j|jg}|dkrDdd|D}t|}t|||||||ddS)audtype-aware variant of numpy.testing.assert_allclose This variant introspects the least precise floating point dtype in the input argument and automatically sets the relative tolerance parameter to 1e-4 float32 and use 1e-7 otherwise (typically float64 in scikit-learn). `atol` is always left to 0. by default. It should be adjusted manually to an assertion-specific value in case there are null values expected in `desired`. The aggregate tolerance is `atol + rtol * abs(desired)`. Parameters ---------- actual : array_like Array obtained. desired : array_like Array desired. rtol : float, optional, default=None Relative tolerance. If None, it is set based on the provided arrays' dtypes. atol : float, optional, default=0. Absolute tolerance. equal_nan : bool, optional, default=True If True, NaNs will compare equal. err_msg : str, optional, default='' The error message to be printed in case of failure. verbose : bool, optional, default=True If True, the conflicting values are appended to the error message. Raises ------ AssertionError If actual and desired are not equal up to specified precision. See Also -------- numpy.testing.assert_allclose Examples -------- >>> import numpy as np >>> from sklearn.utils._testing import assert_allclose >>> x = [1e-5, 1e-3, 1e-1] >>> y = np.arccos(np.cos(x)) >>> assert_allclose(x, y, rtol=1e-5, atol=0) >>> a = np.full(shape=10, fill_value=1e-5, dtype=np.float32) >>> assert_allclose(a, 1e-5) NcSsg|]}|tjkrdndqS)g-C6?Hz>)r%Zfloat32)r'dtyper#r#r)r*?sz#assert_allclose..)rtolatol equal_nanerr_msgverbose)r%Z asanyarrayrnmaxnp_assert_allclose) actualZdesiredrorprqrrrsZdtypesZrtolsr#r#r)r s5 rm& .>cCstj|rttj|rt|}|}||t|j|j|dt|j|j|dt|j |j |||dn4tj|stj|st|||||dnt ddS)aAssert allclose for sparse and dense data. Both x and y need to be either sparse or dense, they can't be mixed. Parameters ---------- x : {array-like, sparse matrix} First array to compare. y : {array-like, sparse matrix} Second array to compare. rtol : float, default=1e-07 relative tolerance; see numpy.allclose. atol : float, default=1e-9 absolute tolerance; see numpy.allclose. Note that the default here is more tolerant than the default for numpy.testing.assert_allclose, where atol=0. err_msg : str, default='' Error message to raise. )rr)rorprrzGCan only compare two sparse matrices, not a sparse matrix and an array.N) spsparseissparseZtocsrZsum_duplicatesrindicesZindptrr datarC)xyrorprrr#r#r)assert_allclose_dense_sparseMsrcCsd|kr|j|ddS)a~Set random state of an estimator if it has the `random_state` param. Parameters ---------- estimator : object The estimator. random_state : int, RandomState instance or None, default=0 Pseudo random number generator state. Pass an int for reproducible results across multiple function calls. See :term:`Glossary `. random_state)rN) get_params set_params)Z estimatorrr#r#r)set_random_statews rzskipped on 32bit platforms)reasonZTRAVIStruezskip on travisznot compatible with PyPyz+OpenBLAS is unstable for this configurationzjoblib is in serial modedarwinz)Possible multi-process bug with some BLAScCsttjddrtddS)NZSKLEARN_SKIP_NETWORK_TESTSrz-Text tutorial requires large dataset download)intosenvirongetrr#r#r#r)check_skip_networksrFcCsFztj|rt|Wn&tk r@|r.c s,g|]$\}}t|td|dqS)r|z.dat)ropr7)r'indexrfolderrr#r)r*s z8_create_aligned_memmap_backed_arrays..z`When creating aligned memmap-backed arrays, input must be a single array or a sequence of arrays) r?r%rrr7rrall enumeraterC)r|rrrr#rr)$_create_aligned_memmap_backed_arrayss    rrc Cstjdd}ttjt|ddtddtD}|r>d}|rPt |||}n&t |d}t ||t j||d}|s~|n||f}|S) aq Parameters ---------- data mmap_mode : str, default='r' return_folder : bool, default=False aligned : bool, default=False If True, if input is a single numpy array and if the input array is aligned, the memory mapped array will also be aligned. This is a workaround for https://github.com/joblib/joblib/issues/563. Zsklearn_testing_)prefixT)rcss2|]*}|ddkr|dddkrdVqdS)Z internal_apiZopenblas architectureZprescottTN)rlower)r'infor#r#r)r/ s z,create_memmap_backed_data..zdata.pkl)r)tempfilemkdtempatexitregister functoolspartialranyrrrr7joblibdumpload) r|rrZalignedrZhas_prescott_openblasZmemmap_backed_datarr<r#r#r)rs&   rcCspzt|j}Wntk r&gYSXdd|D}|rhdd|D}t|dkr`d}||fS|SdS)z!Helper to get function arguments.cSs&g|]\}}|j|j|jfkr|qSr#)kindVAR_POSITIONAL VAR_KEYWORD)r'keyparamr#r#r)r*/sz_get_args..cSsg|]}|j|jkr|jqSr#)rrrW)r'rr#r#r)r*5s rN)r parametersrCitemsvaluesr4)rgvarargsparamsr9r#r#r) _get_args's  rcCs\g}t|}|r||j|j}||jkrF||d|d||jd|S)zGet function full name. Parameters ---------- func : callable The function object. Returns ------- name : str The function name. N.)inspect getmodulerVr6rbfindr7)r8partsmodulequalnamer#r#r)_get_func_nameAs     rc sddlm}g}dkrgnt|}|dr<|dr@|St|rN|S|dddkrd|S|dd d krz|Sttfd d t |}t |dkr|dd kr| d |dkrg}t j ddt dtz||}Wntk rn}zbdt|krPt|ddd} |d|g| 7}|WY W5QRS|t|W5d}~XYnNtk r}z.||dt|g7}|WYW5QRSd}~XYnXW5QRXt |rtd||dfg} |dD]\} } } | shd| krB| d| dddrB||d| g7}n&| drh||d| g7}d| kr| | dddqt |dkr|Sttfdd | } g} ttt | t |D]:}||| |kr| d|||| |fg7} qqt |t | kr8| d|t | g7} n(t |t | kr`| d | t |g7} t | dkrrgSddl}ddl}|| }||}| d!g7} | d"d#|!||D| | d|g|}|S)$arHelper to check docstring. Parameters ---------- func : callable The function object to test. doc : str, default=None Docstring if it is passed manually to the test. ignore : list, default=None Parameters to ignore. Returns ------- incorrect : list A list of string describing the incorrect results. r) docscrapeNzsklearn.zsklearn.externalsr)Z setup_moduleZteardown_moduleZestimator_checkscs|kSr,r#r}rPr#r)z,check_docstring_parameters..rOTrerrorz"potentially wrong underline length z In function: z parsing error: zError for %s: %sZ Parameters:z9 There was no space between the param name and colon (%r)z6 Parameter %r has an empty type spec. Remove the colon*z` cs|kSr,r#rrr#r)rrzmThere's a parameter name mismatch in function docstring w.r.t. function signature, at index %s diff: %r != %rzbParameters in function docstring have less items w.r.t. function signature, first missing item: %sz`Parameters in function docstring have more items w.r.t. function signature, first extra item: %sz Full diff:css|]}|VqdSr,)strip)r'liner#r#r)r/sz-check_docstring_parameters..)"Znumpydocrr startswithrisdatadescriptorsplitlistfilterrr4remover0r1r2 UserWarningZ FunctionDocr-rV ExceptionrYrrrstripendswithlstriprangemindifflibpprintpformat splitlinesextendZndiff)r8docrPrZ incorrect func_nameZparam_signaturerecordsexprfZ param_docsrWZtype_definitionZ param_docirrZparam_docs_formattedZparam_signature_formattedr#rr)check_docstring_parameters[s     0  *         r<c Cstjdd\}}t|zbt|d}||dW5QRXtj |g}t t t tjd}tj}ztj ||dg|d<Wntk r||d<YnX|t|d}tjd} | r| |d d<||d <zfzt|f|} Wn8tk r&} ztd | jdW5d } ~ XYnX| d kr@t| dWn8tk rz} ztd| jdW5d } ~ XYnXW5t|Xd S)aUtility to check assertions in an independent Python subprocess. The script provided in the source code should return 0 and not print anything on stderr or stdout. This is a port from cloudpickle https://github.com/cloudpipe/cloudpickle Parameters ---------- source_code : str The Python source code to execute. timeout : int, default=60 Time in seconds before timeout. z_src_test_sklearn.py)suffixwbzutf-8z.. PYTHONPATH)cwdstderrenvZCOVERAGE_PROCESS_STARTrtimeoutzscript errored with output: %sNrz!script timeout, output so far: %s)rmkstemprcloseunlinkopenwriteencoderJ executablernormpathr7dirnamesklearn__file__rcopypathsepKeyErrorrrrrrYoutputdecoder5r) Z source_coderfdZ source_filefcmdrrrQZ coverage_rcoutr(r#r#r)rs>        cCsB|dkr.|dkrt|Stj||dSn|dkr^|dkrFt|Sttj||dSn|dkrttj||dS|dkrtjj||dS|dkrt d}|j |||d S|d krt d}|j ||dS|d krt d}|j ||dS|d kr t |d |dS|dkr$tjj||dS|dkr>tjj||dSdS)aConvert a given container to a specific array-like with a dtype. Parameters ---------- container : array-like The container to convert. constructor_name : {"list", "tuple", "array", "sparse", "dataframe", "series", "index", "slice", "sparse_csr", "sparse_csc"} The type of the returned container. columns_name : index or array-like, default=None For pandas container supporting `columns_names`, it will affect specific names. dtype : dtype, default=None Force the dtype of the container. Does not apply to `"slice"` container. Returns ------- converted_container rN)rnrdrryZ dataframeZpandas)columnsrnZseriesrslicerZ sparse_csrZ sparse_csc)rr%ZasarraytolistrdrxryZ csr_matrixpytestZ importorskipZ DataFrameZSeriesZIndexrZ csc_matrix) containerZconstructor_nameZ columns_namernpdr#r#r)_convert_container!s6      rcCst||||S)a[Context manager to ensure exceptions are raised within a code block. This is similar to and inspired from pytest.raises, but supports a few other cases. This is only intended to be used in estimator_checks.py where we don't want to use pytest. In the rest of the code base, just use pytest.raises instead. Parameters ---------- excepted_exc_type : Exception or list of Exception The exception that should be raised by the block. If a list, the block should raise one of the exceptions. match : str or list of str, default=None A regex that the exception message should match. If a list, one of the entries must match. If None, match isn't enforced. may_pass : bool, default=False If True, the block is allowed to not raise an exception. Useful in cases where some estimators may support a feature but others must fail with an appropriate error message. By default, the context manager will raise an exception if the block does not raise an exception. err_msg : str, default=None If the context manager fails (e.g. the block fails to raise the proper exception, or fails to match), then an AssertionError is raised with this message. By default, an AssertionError is raised with a default error message (depends on the kind of failure). Use this to indicate how users should fix their estimators to pass the checks. Attributes ---------- raised_and_matched : bool True if an exception was raised and a match was found, False otherwise. )_Raises)expected_exc_typematchmay_passrrr#r#r)raisesUs%rc@seZdZddZddZdS)rcCsBt|tr|n|g|_t|tr&|gn||_||_||_d|_dS)NF)r?rexpected_exc_typesr-matchesrrrraised_and_matched)rOrrrrrr#r#r)rsz_Raises.__init__csdkr,|jrdS|jp"d|j}t|tfdd|jDs`|jdk r\t|jndS|jdk r|jpdd|jt}tfdd|jDst|d|_ dS) NTzDid not raise: c3s|]}t|VqdSr,)rA)r'Z expected_type)rr#r)r/sz#_Raises.__exit__..FzIThe error message should contain one of the following patterns: {} Got {}rc3s|]}t|tVqdSr,)researchr-)r'r) exc_valuer#r)r/s) rrrrr5rrrDr7r-r)rOrr _rrr#)rr r)r`s,     z_Raises.__exit__N)r6rarbrr`r#r#r#r)r}s rc@sPeZdZdZdZdddZdddZd d Zd d Zd dZ ddZ ddZ dS)MinimalClassifiera Minimal classifier implementation with inheriting from BaseEstimator. This estimator should be tested with: * `check_estimator` in `test_estimator_checks.py`; * within a `Pipeline` in `test_pipeline.py`; * within a `SearchCV` in `test_search.py`. classifierNcCs ||_dSr,rrOrr#r#r)rszMinimalClassifier.__init__TcCs d|jiSNrr$rOdeepr#r#r)rszMinimalClassifier.get_paramscKs"|D]\}}t|||q|Sr,rsetattrrOrrvaluer#r#r)rszMinimalClassifier.set_paramscCs8t||\}}t|tj|dd\|_}||_|S)NT)Z return_counts)rrr%uniqueclasses_argmax_most_frequent_class_idx)rOXr~countsr#r#r)fits  zMinimalClassifier.fitcCsHt|t|}|jd|jjf}tj|tjd}d|dd|jf<|S)Nr)rrng?) rrrr.sizer%zerosZfloat64r0)rOr1Z proba_shapey_probar#r#r) predict_probas zMinimalClassifier.predict_probacCs ||}|jdd}|j|S)Nr)Zaxis)r7r/r.)rOr1r6Zy_predr#r#r)predicts  zMinimalClassifier.predictcCsddlm}||||S)Nr)accuracy_score)sklearn.metricsr9r8)rOr1r~r9r#r#r)scores zMinimalClassifier.score)N)T) r6rarbrc_estimator_typerrrr3r7r8r;r#r#r#r)r"s   r"c@sHeZdZdZdZdddZdddZd d Zd d Zd dZ ddZ dS)MinimalRegressora Minimal regressor implementation with inheriting from BaseEstimator. This estimator should be tested with: * `check_estimator` in `test_estimator_checks.py`; * within a `Pipeline` in `test_pipeline.py`; * within a `SearchCV` in `test_search.py`. Z regressorNcCs ||_dSr,r$r%r#r#r)rszMinimalRegressor.__init__TcCs d|jiSr&r$r'r#r#r)rszMinimalRegressor.get_paramscKs"|D]\}}t|||q|Sr,r)r+r#r#r)rszMinimalRegressor.set_paramscCs$t||\}}d|_t||_|SNT)r is_fitted_r%Zmean_meanrOr1r~r#r#r)r3s zMinimalRegressor.fitcCs*t|t|}tj|jdfd|jS)Nr)r)rrr%Zonesrr@)rOr1r#r#r)r8szMinimalRegressor.predictcCsddlm}||||S)Nr)r2_score)r:rBr8)rOr1r~rBr#r#r)r;s zMinimalRegressor.score)N)T) r6rarbrcr<rrrr3r8r;r#r#r#r)r=s   r=c@sJeZdZdZdddZdddZdd Zdd d Zdd d ZdddZ dS)MinimalTransformeraMinimal transformer implementation with inheriting from BaseEstimator. This estimator should be tested with: * `check_estimator` in `test_estimator_checks.py`; * within a `Pipeline` in `test_pipeline.py`; * within a `SearchCV` in `test_search.py`. NcCs ||_dSr,r$r%r#r#r)rszMinimalTransformer.__init__TcCs d|jiSr&r$r'r#r#r)rszMinimalTransformer.get_paramscKs"|D]\}}t|||q|Sr,r)r+r#r#r)rszMinimalTransformer.set_paramscCst|d|_|Sr>)rr?rAr#r#r)r3szMinimalTransformer.fitcCst|t|}|Sr,)rrrAr#r#r) transformszMinimalTransformer.transformcCs|||||Sr,)r3rDrAr#r#r) fit_transform#sz MinimalTransformer.fit_transform)N)T)N)N)N) r6rarbrcrrrr3rDrEr#r#r#r)rCs    rC)NrkTrlT)rmrwrl)r)F)rFF)F)NN)r)NN)NFN)ircros.pathrrrr0rJrr subprocessrrrrr contextlibcollections.abcrrZscipyrxr r rrZunittestr r NameErrorZ numpy.testingr rur rrrrZnumpyr%rrZ sklearn.utilsrrrZsklearn.utils.multiclassrZsklearn.utils.validationrrrZsklearn.utils.fixesr__all__Z_dummy assertRaisesrZcaserZassertDictEqualZassert_dict_equalassertRaisesRegexZassert_raises_regexrr=rBrHrFrjrrrmarkZskipifZ skip_if_32bitrrZ skip_travisZxfailZ fails_if_pypyZfails_if_unstable_openblasparallelmpZskip_if_no_parallelplatformZ!if_safe_multiprocessing_with_blas ImportErrorrrrrrrrrrrrrAbstractContextManagerrr"r=rCr#r#r#r)s                  ,9+ I *     ,   3 4 (-1(