自然语言处理 | 使用Spacy 进行自然语言处理（二）

上次我们简单介绍了Spacy，学习了它的安装以及实体识别等基本的方法。今天我继续给大家介绍一下它的其他功能如何操作，主要有词性还原，词性标注，名词块识别，依存分析等内容。废话不多说，直接看代码。

1. import en_core_web_sm

3. parser = en_core_web_sm.load()

4. sentences = "There is an art, it says, or rather, a knack to flying." \

5.            "The knack lies in learning how to throw yourself at the ground and miss." \

6.            "In the beginning the Universe was created. This has made a lot of people " \

7.            "very angry and been widely regarded as a bad move."

9. print("解析文本中包含的句子：")

10. sents = [sent for sent in parser(sentences).sents]

11. for x in sents:

12.    print(x)

13. """

14. There is an art, it says, or rather, a knack to flying.

15. The knack lies in learning how to throw yourself at the ground and miss.

16. In the beginning the Universe was created.

17. This has made a lot of people very angry and been widely regarded as a bad move.

18. """

19. print("- * -"*20)

21. # 分词

22. print()

23. tokens = [token for token in sents[0] if len(token) > 1]

24. print(tokens)

25. print("- * -"*20)

27. # 词性还原

28. lemma_tokens = [token.lemma_ for token in sents[0] if len(token) > 1]

30. print(lemma_tokens)

31. print("- * -"*20)

33. # 简化版的词性标注

34. pos_tokens = [token.pos_ for token in sents[0] if len(token) > 1]

35. print(pos_tokens)

36. print("- * -"*20)

38. # 词性标注的细节版

39. tag_tokens = [token.tag_ for token in sents[0] if len(token) > 1]

40. print(tag_tokens)

41. print("- * -"*20)

43. # 依存分析

44. dep_tokens = [token.dep_ for token in sents[0] if len(token) > 1]

45. print(dep_tokens)

46. print("- * -"*20)

49. print("名词块分析")

50. doc = parser(u"Autonomous cars shift insurance liability toward manufacturers")

51. # 获取名词块文本

52. chunk_text = [chunk.text for chunk in doc.noun_chunks]

53. print(chunk_text)

54. print("- * -"*20)

56. # 获取名词块根结点的文本

57. chunk_root_text = [chunk.root.text for chunk in doc.noun_chunks]

58. print(chunk_root_text)

59. print("- * -"*20)

61. # 依存分析

62. chunk_root_dep_ = [chunk.root.dep_ for chunk in doc.noun_chunks]

63. print(chunk_root_dep_)

64. print("- * -"*20)

65. #

66. chunk_root_head_text = [chunk.root.head.text for chunk in doc.noun_chunks]

67. print(chunk_root_head_text)

68. print("- * -"*20)

最后给大家附上一个句法依存分析的结果解释的资料，是斯坦福自然语言处理的一个依存句法分析的解释文档

链接：https://nlp.stanford.edu/software/dependencies_manual.pdf

如果下载不下来，可以微信和我要。

百度文库有中文版：https://wenku.baidu.com/view/1e92891dbceb19e8b8f6bae5.html