Mercurial > hg > ooxml

annotate excel.py @ 71:54bb53434887

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
begin work on decoder that allows identifiers as keys and values
author Henry S. Thompson <ht@markup.co.uk>
date Wed, 28 Jun 2017 19:08:30 +0100
parents 0003fe7b6b67
children 4bd5de7ac247
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
70
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
1 #!/usr/bin/python3
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
2 '''Class model for analysis of Excel spreadsheets'''
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
3 from jsonweb.encode import to_object, dumper
71
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
4 from jsonweb.decode import from_object, loader
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
5 import json
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
6 import eDecoder
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
7
70
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
8 try:
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
9 string_types=basestring
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
10 except NameError:
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
11 string_types=str
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
12
71
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
13 @from_object()
70
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
14 @to_object(exclude_nulls=True)
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
15 class Book(object):
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
16 def __init__(self,source,sheets=[],formats=[]):
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
17 assert(isinstance(source,string_types))
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
18 self.source=source
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
19 sheets=list(sheets)
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
20 assert(all(isinstance(s,Sheet) for s in sheets))
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
21 self.sheets=sheets
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
22 formats=list(formats)
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
23 self.formats=formats
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
24
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
25 def addSheet(self,sheet):
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
26 assert(isinstance(sheet,Sheet))
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
27 self.sheets.append(sheet)
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
28
71
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
29 @from_object()
70
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
30 @to_object(exclude_nulls=True,suppress=["book"])
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
31 class Sheet(object):
71
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
32 def __init__(self,name,book=None,tables=[],docs=[],misc=[]):
70
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
33 assert(isinstance(name,string_types))
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
34 self.name=name
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
35 tables=list(tables)
71
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
36 assert(all(isinstance(s,Table) for s in tables))
70
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
37 self.tables=tables
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
38 docs=list(docs)
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
39 assert(all(isinstance(s,Region) for s in docs))
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
40 self.docs=docs
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
41 misc=list(misc)
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
42 assert(all(isinstance(s,Region) for s in misc))
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
43 self.misc=misc
71
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
44 if book is not None:
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
45 assert(isinstance(book,Book))
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
46 book.addSheet(self)
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
47 self.book=book
70
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
48
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
49 def addTable(self,table):
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
50 assert(isinstance(table,Region))
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
51 self.tables.append(table)
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
52
71
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
53 @from_object()
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
54 @to_object(exclude_nulls=True,suppress=["parent"])
70
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
55 class Region(object):
71
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
56 def __init__(self,name,parent=None,ranges=[],content=[]):
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
57 if parent is not None:
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
58 assert(isinstance(parent,(Region,Sheet)))
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
59 self.parent=parent
70
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
60 ranges=list(ranges)
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
61 assert(all(isinstance(s,Range) for s in ranges))
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
62 self.ranges=ranges
71
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
63 content=list(content)
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
64 assert(all(isinstance(s,(Region,string_types)) for s in content))
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
65 self.content=content
70
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
66 assert(isinstance(name,string_types))
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
67 self.name=name
71
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
68 if parent is not None:
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
69 parent.addRegion(self)
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
70
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
71 def addRegion(self,content):
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
72 assert(isinstance(content,(Region,string_types)))
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
73 self.content.append(content)
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
74
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
75 @from_object()
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
76 @to_object(exclude_nulls=True,suppress=["sheet"])
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
77 class Table(Region):
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
78 def __init__(self,name,sheet=None,ranges=[],labels=[],data=[]):
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
79 super(Table,self).__init__(name,sheet,ranges)
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
80 labels=list(labels)
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
81 assert(all(isinstance(s,Label) for s in labels))
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
82 self.labels=labels
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
83 data=list(data)
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
84 assert(all(isinstance(s,Region) for s in data))
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
85 self.data=data
70
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
86 if sheet is not None:
71
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
87 assert(isinstance(sheet,Sheet))
70
0003fe7b6b67 beginning work on class structure for excel annotation
Henry S. Thompson <ht@markup.co.uk>
parents:
diff changeset
88 sheet.addTable(self)
71
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
89 self.sheet=sheet
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
90
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
91 def addLabel(self,label):
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
92 assert(isinstance(label,Label))
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
93 self.labels.append(label)
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
94
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
95 def addData(self,data):
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
96 assert(isinstance(data,Data))
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
97 self.data.append(data)
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
98
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
99 def addRegion(self,region):
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
100 assert(isinstance(region,(Label,Data)))
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
101 if isinstance(region,Label):
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
102 self.addLabel(region)
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
103 else:
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
104 self.addData(region)
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
105
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
106 def lt(filename):
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
107 with open(filename,'r') as js:
54bb53434887 begin work on decoder that allows identifiers as keys and values
Henry S. Thompson <ht@markup.co.uk>
parents: 70
diff changeset
108 return json.loads(js.read(),cls=eDecoder.eDecoder)