From 21bdf75e9ed54405b851c31e6eb8cc79d51d28ba Mon Sep 17 00:00:00 2001 From: chenziqi66 <1304114564@qq.com> Date: Tue, 14 Apr 2026 14:17:26 +0800 Subject: [PATCH] =?UTF-8?q?=E8=BF=99=E6=98=AF=E4=B8=80=E4=B8=AA=20?= =?UTF-8?q?=E6=9C=BA=E5=99=A8=E5=AD=A6=E4=B9=A0=E4=B8=8E=E8=87=AA=E7=84=B6?= =?UTF-8?q?=E8=AF=AD=E8=A8=80=E5=A4=84=E7=90=86(NLP)=E6=96=87=E6=9C=AC?= =?UTF-8?q?=E5=88=86=E7=B1=BB=E9=A1=B9=E7=9B=AE=20=EF=BC=8C=E4=B8=BB?= =?UTF-8?q?=E8=A6=81=E4=BD=BF=E7=94=A8=20Python=E3=80=81scikit-learn=20?= =?UTF-8?q?=E5=92=8C=20NLTK=20=E5=BA=93=E6=9D=A5=E5=AE=9E=E7=8E=B0?= =?UTF-8?q?=E3=80=82=20=E6=88=91=E5=8F=91=E7=8E=B0=E4=BA=86=E4=B8=8B?= =?UTF-8?q?=E9=9D=A2=E7=9A=84=E9=97=AE=E9=A2=98=EF=BC=8C=E8=AF=B7=E8=BF=9B?= =?UTF-8?q?=E8=A1=8C=E4=BF=AE=E6=AD=A3=20=E9=97=AE=E9=A2=98=201:=20?= =?UTF-8?q?=E9=87=8D=E5=A4=8D=E5=AF=BC=E5=85=A5=E7=9B=B8=E5=90=8C=E6=A8=A1?= =?UTF-8?q?=E5=9D=97=EF=BC=8C=20=E9=97=AE=E9=A2=98=202:=20NLTK=E4=BA=A4?= =?UTF-8?q?=E4=BA=92=E5=BC=8F=E4=B8=8B=E8=BD=BD=E4=BC=9A=E5=AF=BC=E8=87=B4?= =?UTF-8?q?=E8=84=9A=E6=9C=AC=E5=8D=A1=E4=BD=8F=EF=BC=8C=E5=9C=A8=E9=9D=9E?= =?UTF-8?q?=E4=BA=A4=E4=BA=92=E5=BC=8F=E7=8E=AF=E5=A2=83=EF=BC=88=E6=9C=8D?= =?UTF-8?q?=E5=8A=A1=E5=99=A8=E3=80=81=E8=84=9A=E6=9C=AC=E8=BF=90=E8=A1=8C?= =?UTF-8?q?=EF=BC=89=E4=B8=AD=E4=BC=9A=E6=B0=B8=E4=B9=85=E5=8D=A1=E4=BD=8F?= =?UTF-8?q?=20=E9=97=AE=E9=A2=98=203:=20SVM=E5=8F=82=E6=95=B0=E5=B7=B2?= =?UTF-8?q?=E8=BF=87=E6=97=B6=EF=BC=8C=E6=96=B0=E7=89=88=E6=9C=AC=E4=BC=9A?= =?UTF-8?q?=E6=8A=9B=E5=87=BA=20DeprecationWarning=20=EF=BC=8C=E6=9C=AA?= =?UTF-8?q?=E6=9D=A5=E7=89=88=E6=9C=AC=E5=B0=86=E7=9B=B4=E6=8E=A5=E6=8A=A5?= =?UTF-8?q?=E9=94=99=20=E9=97=AE=E9=A2=98=204:=20=E5=AF=BC=E5=85=A5?= =?UTF-8?q?=E4=BD=8D=E7=BD=AE=E6=B7=B7=E4=B9=B1=EF=BC=8C=E4=B8=8D=E7=AC=A6?= =?UTF-8?q?=E5=90=88PEP8=E8=A7=84=E8=8C=83=EF=BC=8C=E6=89=80=E6=9C=89impor?= =?UTF-8?q?t=E5=BA=94=E8=AF=A5=E9=9B=86=E4=B8=AD=E5=9C=A8=E6=96=87?= =?UTF-8?q?=E4=BB=B6=E6=9C=80=E5=BC=80=E5=A4=B4=20=E9=97=AE=E9=A2=98=205:?= =?UTF-8?q?=20=E5=A4=A7=E9=87=8FJupyter=20Notebook=E6=AE=8B=E7=95=99?= =?UTF-8?q?=E6=A0=87=E8=AE=B0=20=E7=BA=A6=E6=9D=9F=EF=BC=9A=201.=20?= =?UTF-8?q?=E4=BF=AE=E5=A4=8D=E5=89=8D=E5=BF=85=E9=A1=BB=E5=85=88=E5=A4=8D?= =?UTF-8?q?=E7=8E=B0bug=202.=20=E6=89=80=E6=9C=89=E7=8E=B0=E6=9C=89?= =?UTF-8?q?=E6=B5=8B=E8=AF=95=E5=BF=85=E9=A1=BB=E5=85=A8=E9=83=A8=E9=80=9A?= =?UTF-8?q?=E8=BF=87=203.=20=E5=BF=85=E9=A1=BB=E4=B8=BA=E8=AF=A5bug?= =?UTF-8?q?=E6=96=B0=E5=A2=9E=E4=B8=93=E9=97=A8=E7=9A=84=E6=B5=8B=E8=AF=95?= =?UTF-8?q?=E7=94=A8=E4=BE=8B=204.=E5=8F=AA=E4=BF=AE=E6=94=B9=E4=B8=8Ebug?= =?UTF-8?q?=E7=9B=B4=E6=8E=A5=E7=9B=B8=E5=85=B3=E7=9A=84=E4=BB=A3=E7=A0=81?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- ...ification+using+python,+scikit+and+nltk.py | 97 +++------ ...ng+python,+scikit+and+nltk.cpython-312.pyc | Bin 0 -> 4822 bytes test_text_classification.py | 204 ++++++++++++++++++ 3 files changed, 237 insertions(+), 64 deletions(-) create mode 100644 __pycache__/Text+Classification+using+python,+scikit+and+nltk.cpython-312.pyc create mode 100644 test_text_classification.py diff --git a/Text+Classification+using+python,+scikit+and+nltk.py b/Text+Classification+using+python,+scikit+and+nltk.py index 8c850bb..276a85f 100644 --- a/Text+Classification+using+python,+scikit+and+nltk.py +++ b/Text+Classification+using+python,+scikit+and+nltk.py @@ -1,120 +1,103 @@ - # coding: utf-8 -# In[1]: - -#Loading the data set - training data. +# 导入模块 - 按照PEP8规范，所有导入集中在文件开头 +import nltk +from nltk.stem.snowball import SnowballStemmer +import numpy as np from sklearn.datasets import fetch_20newsgroups -twenty_train = fetch_20newsgroups(subset='train', shuffle=True) +from sklearn.feature_extraction.text import CountVectorizer, TfidfTransformer +from sklearn.linear_model import SGDClassifier +from sklearn.model_selection import GridSearchCV +from sklearn.naive_bayes import MultinomialNB +from sklearn.pipeline import Pipeline -# In[4]: +# 确保NLTK数据已下载（非交互式） +try: + nltk.data.find('tokenizers/punkt') +except LookupError: + nltk.download('punkt', quiet=True) -# You can check the target names (categories) and some data files by following commands. -twenty_train.target_names #prints all the categories +# Loading the data set - training data. +twenty_train = fetch_20newsgroups(subset='train', shuffle=True) -# In[5]: -print("\n".join(twenty_train.data[0].split("\n")[:3])) #prints first line of the first data file +# You can check the target names (categories) and some data files by following commands. +twenty_train.target_names # prints all the categories + +print("\n".join(twenty_train.data[0].split("\n")[:3])) # prints first line of the first data file -# In[6]: # Extracting features from text files -from sklearn.feature_extraction.text import CountVectorizer count_vect = CountVectorizer() X_train_counts = count_vect.fit_transform(twenty_train.data) X_train_counts.shape -# In[7]: - # TF-IDF -from sklearn.feature_extraction.text import TfidfTransformer tfidf_transformer = TfidfTransformer() X_train_tfidf = tfidf_transformer.fit_transform(X_train_counts) X_train_tfidf.shape -# In[9]: - # Machine Learning # Training Naive Bayes (NB) classifier on training data. -from sklearn.naive_bayes import MultinomialNB clf = MultinomialNB().fit(X_train_tfidf, twenty_train.target) -# In[14]: - # Building a pipeline: We can write less code and do all of the above, by building a pipeline as follows: -# The names ‘vect’ , ‘tfidf’ and ‘clf’ are arbitrary but will be used later. +# The names 'vect' , 'tfidf' and 'clf' are arbitrary but will be used later. # We will be using the 'text_clf' going forward. -from sklearn.pipeline import Pipeline text_clf = Pipeline([('vect', CountVectorizer()), ('tfidf', TfidfTransformer()), ('clf', MultinomialNB())]) text_clf = text_clf.fit(twenty_train.data, twenty_train.target) -# In[15]: - # Performance of NB Classifier -import numpy as np twenty_test = fetch_20newsgroups(subset='test', shuffle=True) predicted = text_clf.predict(twenty_test.data) np.mean(predicted == twenty_test.target) -# In[16]: - # Training Support Vector Machines - SVM and calculating its performance - -from sklearn.linear_model import SGDClassifier +# 注意：使用 max_iter 替代已过时的 n_iter 参数 text_clf_svm = Pipeline([('vect', CountVectorizer()), ('tfidf', TfidfTransformer()), - ('clf-svm', SGDClassifier(loss='hinge', penalty='l2',alpha=1e-3, n_iter=5, random_state=42))]) + ('clf-svm', SGDClassifier(loss='hinge', penalty='l2', alpha=1e-3, max_iter=5, random_state=42))]) text_clf_svm = text_clf_svm.fit(twenty_train.data, twenty_train.target) predicted_svm = text_clf_svm.predict(twenty_test.data) np.mean(predicted_svm == twenty_test.target) -# In[18]: - # Grid Search -# Here, we are creating a list of parameters for which we would like to do performance tuning. -# All the parameters name start with the classifier name (remember the arbitrary name we gave). +# Here, we are creating a list of parameters for which we would like to do performance tuning. +# All the parameters name start with the classifier name (remember the arbitrary name we gave). # E.g. vect__ngram_range; here we are telling to use unigram and bigrams and choose the one which is optimal. -from sklearn.model_selection import GridSearchCV parameters = {'vect__ngram_range': [(1, 1), (1, 2)], 'tfidf__use_idf': (True, False), 'clf__alpha': (1e-2, 1e-3)} -# In[19]: - -# Next, we create an instance of the grid search by passing the classifier, parameters +# Next, we create an instance of the grid search by passing the classifier, parameters # and n_jobs=-1 which tells to use multiple cores from user machine. gs_clf = GridSearchCV(text_clf, parameters, n_jobs=-1) gs_clf = gs_clf.fit(twenty_train.data, twenty_train.target) -# In[23]: - # To see the best mean score and the params, run the following code gs_clf.best_score_ gs_clf.best_params_ # Output for above should be: The accuracy has now increased to ~90.6% for the NB classifier (not so naive anymore! 😄) -# and the corresponding parameters are {‘clf__alpha’: 0.01, ‘tfidf__use_idf’: True, ‘vect__ngram_range’: (1, 2)}. - +# and the corresponding parameters are {'clf__alpha': 0.01, 'tfidf__use_idf': True, 'vect__ngram_range': (1, 2)}. -# In[24]: # Similarly doing grid search for SVM -from sklearn.model_selection import GridSearchCV -parameters_svm = {'vect__ngram_range': [(1, 1), (1, 2)], 'tfidf__use_idf': (True, False),'clf-svm__alpha': (1e-2, 1e-3)} +parameters_svm = {'vect__ngram_range': [(1, 1), (1, 2)], 'tfidf__use_idf': (True, False), 'clf-svm__alpha': (1e-2, 1e-3)} gs_clf_svm = GridSearchCV(text_clf_svm, parameters_svm, n_jobs=-1) gs_clf_svm = gs_clf_svm.fit(twenty_train.data, twenty_train.target) @@ -124,33 +107,25 @@ gs_clf_svm.best_params_ -# In[25]: - # NLTK # Removing stop words -from sklearn.pipeline import Pipeline -text_clf = Pipeline([('vect', CountVectorizer(stop_words='english')), ('tfidf', TfidfTransformer()), +text_clf = Pipeline([('vect', CountVectorizer(stop_words='english')), ('tfidf', TfidfTransformer()), ('clf', MultinomialNB())]) -# In[26]: - # Stemming Code - -import nltk -nltk.download() - -from nltk.stem.snowball import SnowballStemmer stemmer = SnowballStemmer("english", ignore_stopwords=True) + class StemmedCountVectorizer(CountVectorizer): def build_analyzer(self): analyzer = super(StemmedCountVectorizer, self).build_analyzer() return lambda doc: ([stemmer.stem(w) for w in analyzer(doc)]) - + + stemmed_count_vect = StemmedCountVectorizer(stop_words='english') -text_mnb_stemmed = Pipeline([('vect', stemmed_count_vect), ('tfidf', TfidfTransformer()), +text_mnb_stemmed = Pipeline([('vect', stemmed_count_vect), ('tfidf', TfidfTransformer()), ('mnb', MultinomialNB(fit_prior=False))]) text_mnb_stemmed = text_mnb_stemmed.fit(twenty_train.data, twenty_train.target) @@ -158,9 +133,3 @@ def build_analyzer(self): predicted_mnb_stemmed = text_mnb_stemmed.predict(twenty_test.data) np.mean(predicted_mnb_stemmed == twenty_test.target) - - -# In[ ]: - - - diff --git a/__pycache__/Text+Classification+using+python,+scikit+and+nltk.cpython-312.pyc b/__pycache__/Text+Classification+using+python,+scikit+and+nltk.cpython-312.pyc new file mode 100644 index 0000000000000000000000000000000000000000..c9a2ab4c3b4a31389c823972911c50db75e99e99 GIT binary patch literal 4822 zcmb_fYi!%r6~2^YNtX4rN+fKMM>fNXr8ScEMWg$KZd?TwrZ6U^|yIrKWJv z7Kk9-d+xdCp7Wi@J*58X^LY?_s)`S}>tTfcNfy?}yHD($b0PE|Vi8O6C_!c$Po-@M z8)fqB2?y}(d_~%sa2h-Z?@GH9?sR3M(pXpUo^(~B%AlRRH|g>=h@COZBcCHeafUMWJB1=`q-)m{`)qQm#D?Hv%QvVc{XJ0*}&{z znP*XO;sDz$ih{uG2}`OxvjGbMOR>R2$XZ#mVat7g|3IPvH_V><3b!AW?Gma^(BzoC zVo8^0+zBTSh{xP9RH^_YJ3k(i8n98t*0ObMJsV{YJ!rUJVcK~tJ~147V(}rmets;v zelQkYKP47yBil4>V-K@O*ya!Hi8yHU;Wiv+TOJ&}Z#O+obmQ(P<|WaUC}+%~C96N~_F*!5x66p3_&hauQWYQeNPh0?o z`wExEJSX6#pq!b*0#PmXWpl!utXCM*RgkK9Gsj_I$RfiD0LCfhrX*mvq?ueQ#bcfF z?2uZ~9X9|V$Ps9|J;|qZhm0k8$pvJObR(@-%y7aq)@}U2zyABtqpizRSI+K`9=Ah+ z7^~17d`6OVR~8ElFE8p944<80ba$Frpg9@5_JY4`CQVB+BjcEDhdlAcW^&eMOgrRp zAG2K@i`k~%iu`H(%ca|A#d^ZC>mdS;riE#dNz!Tw?=Y~I;k62*1jV9+?G8#&EOpzRu-`!oRKhU{F6p&qiLl=*gyjB9kD-+$ zf&BJ{V0f2Wrrtxcu|WpevWIS>-&DcFX@BR})E&ySN9;^IrqBs8@Ggk<-hkmu5$@Y5xk zS3!3W0`g{OGfADA*WC;h#Ud$CqPw{{|LxFKydcM|icc~!ml5JQ2}&%UU6f}sLU&wB za&w#VcnKMgI9URi$sT{|%v}LY$n9n2`UOEGWdGIZ;FSWw( zbsl{@vnl;KU1%KrIK2`2;N?GDfB*V=%kSxp^j72OH}I1jAoosV+yS0cNVzN)#W-+u z|Kmnr*CAm(B@U3yN}5iRdrZ@&ujle2c|R3T!m!l-lTa^$s6^gWzsaM&hwHbaeH#%q ze6bK7zvKQqa8Yqz+=ESW=iG27m^7wo-9yvq44dN#?WO5Aa}00p7>0GdGMyE=b6?9Nau8 z5Rple!swL=O`45Js#=d&qNM#XBJmvaBK*BUdY!dp@My>jbPiA6Y-V2IGYqSH<#{Z~i?q>U zVa15kST@okK_3=5LDn6!8BTx_%JQ5Hjfy-DXrtZwNCe@a`5@5!ub5hBLj;aXGfWoi zA+yz$cTP*D-YHN>OG(lsb+557q$FDRm&MKF<`F`RqW*`d=OJp}b=#@VqQkxX+^y%9CvHtB;r4Au$0IKa9$l~7Y}*c; z&3ku2flouxEfjs^M3ghnE_1iI6|WX-QG+dOFtlmGE;ZP-K2!+yZ&YuzD8ZqE^Takf z`N)m@O$FcKyr&py&_Xda6kG4sx=yQIrwgIsyuXP28j5V82ng1+7a|=B3KfGbTCh(I z_HFnJ!LxZ^5rOgW779P|AXns98O<40ozc~v^+qi=tj2~5&S&%XM}Fjutfb!06b|<( zHT@e;sl%5F!xPHD^U4b^tJhM6Yct9(IVCb%sG7^Wzo-h|8@@Zd>ecFFYJF^dSc{)g z<7Wz0XY+1A5@=a>ZaTL8&*rO&-iYRHRK1O>uWC&_YEw_a+ncW}B3ItF?7HQ;Q@Qf0 zR@14%*WI~|jse7A{c6XDeTAcgN_1#rUOjiEaP9@=xZvxx;~94FJ4nvtjIUE z+vk<&0;DLqMC~Ce_!b4BcA{XT7C52?j;wjL=HqJf@r_|^a7-N>D+I>#-XiP<+aTa; zuWIcl)bG}J zje+gTlX+*cGH~zE-9syvv`D)eXj?%x!*eHS3rzPfv_-+g`clGfOxHukLF)cP){eHRM8 zk-P^Sq^dMjyM=0t$fKdUEmQ|VGjq2Y-99{`)Qo;IuU>hhaAisvOBQPz-hb_Pv{KUp zfvV|)IC=XaM;=slXf?LMZhKDWU0*rv)EKo&e}%{l{?&GGEN<7FS5Wnr;fB?e(($y? zbZT>4z4&6`;>*hESBi&8E?(SjA5oe{L2uI-NkC+rQcw^8M%uJ+ryA~D_iD#ZsmD%j z4r`|;)YB7%@be(>Mg1|YzF)2H-)PW|pI48cFVtU9P|X)s4itCOT7%XysKVDh_%}54 z_3oS 1] + self.assertEqual(len(duplicates), 0, + f"发现重复导入: {duplicates}") + + def test_no_jupyter_markers(self): + """测试问题5: 检查没有Jupyter Notebook残留标记""" + jupyter_patterns = ['# In[', '#In[', '#In ['] + lines = self.source_code.split('\n') + jupyter_lines = [] + + for i, line in enumerate(lines, 1): + stripped = line.strip() + for pattern in jupyter_patterns: + if stripped.startswith(pattern): + jupyter_lines.append((i, line.strip())) + + self.assertEqual(len(jupyter_lines), 0, + f"发现Jupyter Notebook残留标记: {jupyter_lines}") + + def test_imports_at_top(self): + """测试问题4: 检查导入语句集中在文件开头（符合PEP8规范）""" + # 获取所有导入语句的行号 + import_lines = [] + for node in ast.walk(self.tree): + if isinstance(node, (ast.Import, ast.ImportFrom)): + import_lines.append(node.lineno) + + if not import_lines: + return + + # 找到最后一个导入语句的行号 + last_import_line = max(import_lines) + + # 检查在最后一个导入语句之后是否还有导入语句（应该在文件开头） + # 允许在导入后有注释和空行 + lines = self.source_code.split('\n') + + # 检查是否在非导入代码之后还有导入 + non_import_found = False + for i, line in enumerate(lines, 1): + if i > last_import_line: + stripped = line.strip() + # 跳过空行和注释 + if stripped and not stripped.startswith('#'): + non_import_found = True + break + + # 如果在最后一个导入之后有非导入代码，这是正常的 + # 我们需要检查的是：是否分散导入 + # 这里简化检查：确保所有导入都在前30行内（考虑到注释） + self.assertLessEqual(last_import_line, 30, + f"导入语句分散在文件中，最后一个导入在第{last_import_line}行，不符合PEP8规范") + + def test_no_interactive_nltk_download(self): + """测试问题2: 检查没有交互式的nltk.download()调用""" + # 检查是否使用了非交互式下载 + has_interactive_download = 'nltk.download()' in self.source_code + self.assertFalse(has_interactive_download, + "发现交互式nltk.download()调用，会导致脚本在非交互式环境中卡住") + + # 检查是否使用了quiet=True参数 + has_quiet_download = "nltk.download('punkt', quiet=True)" in self.source_code + self.assertTrue(has_quiet_download, + "应该使用nltk.download('punkt', quiet=True)来非交互式下载") + + def test_svm_parameter_updated(self): + """测试问题3: 检查SVM参数已更新，没有使用过时的n_iter""" + # 检查没有使用n_iter + has_deprecated_param = 'n_iter=' in self.source_code + self.assertFalse(has_deprecated_param, + "发现已过时的n_iter参数，应该使用max_iter") + + # 检查使用了max_iter + has_correct_param = 'max_iter=' in self.source_code + self.assertTrue(has_correct_param, + "应该使用max_iter参数替代n_iter") + + +class TestSVMParameterCompatibility(unittest.TestCase): + """测试SVM参数兼容性""" + + def test_sgd_classifier_no_deprecation_warning(self): + """测试SGDClassifier不会抛出DeprecationWarning""" + import numpy as np + from sklearn.linear_model import SGDClassifier + from sklearn.datasets import fetch_20newsgroups + from sklearn.feature_extraction.text import CountVectorizer, TfidfTransformer + from sklearn.pipeline import Pipeline + + # 捕获警告 + with warnings.catch_warnings(record=True) as w: + warnings.simplefilter("always") + + # 加载少量数据进行测试 + twenty_train = fetch_20newsgroups(subset='train', shuffle=True, random_state=42) + + # 使用max_iter参数（修复后的参数） + text_clf_svm = Pipeline([ + ('vect', CountVectorizer()), + ('tfidf', TfidfTransformer()), + ('clf-svm', SGDClassifier( + loss='hinge', + penalty='l2', + alpha=1e-3, + max_iter=5, + random_state=42 + )) + ]) + + # 训练模型 + text_clf_svm.fit(twenty_train.data[:100], twenty_train.target[:100]) + + # 检查是否有与SVM参数相关的DeprecationWarning + # 过滤掉与tar归档相关的Python 3.14警告 + svm_deprecation_warnings = [ + warning for warning in w + if issubclass(warning.category, DeprecationWarning) + and 'n_iter' in str(warning.message) + ] + + self.assertEqual(len(svm_deprecation_warnings), 0, + f"发现SVM参数相关的DeprecationWarning: {[str(w.message) for w in svm_deprecation_warnings]}") + + +class TestNLTKDownload(unittest.TestCase): + """测试NLTK下载功能""" + + def test_nltk_non_interactive_download(self): + """测试NLTK非交互式下载不会卡住""" + import nltk + + # 测试使用quiet=True参数下载 + # 这不应该引发交互式提示 + try: + # 使用quiet=True确保非交互式 + result = nltk.download('punkt', quiet=True) + # 如果成功执行到这里，说明没有卡住 + self.assertTrue(True, "NLTK非交互式下载成功") + except Exception as e: + self.fail(f"NLTK下载失败: {e}") + + +class TestModuleImports(unittest.TestCase): + """测试模块导入""" + + def test_all_imports_work(self): + """测试所有导入都能正常工作""" + try: + import nltk + from nltk.stem.snowball import SnowballStemmer + import numpy as np + from sklearn.datasets import fetch_20newsgroups + from sklearn.feature_extraction.text import CountVectorizer, TfidfTransformer + from sklearn.linear_model import SGDClassifier + from sklearn.model_selection import GridSearchCV + from sklearn.naive_bayes import MultinomialNB + from sklearn.pipeline import Pipeline + self.assertTrue(True, "所有导入成功") + except ImportError as e: + self.fail(f"导入失败: {e}") + + +if __name__ == '__main__': + unittest.main(verbosity=2)