aboutsummaryrefslogtreecommitdiffstats
path: root/skate/zippy.go
blob: 7ee12cfa092206409038da12ca7576bc177eb2a5 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
// This file contains various "reducers", e.g. working on two data streams and
// applying a function on groups of documents with a shared key.
//
// Note: A bit repetitive, but not want to introduce any other abstraction.
package skate

import (
	"fmt"
	"io"
	"log"
	"strings"

	"git.archive.org/martin/cgraph/skate/zipkey"
	json "github.com/segmentio/encoding/json"
)

// groupLogf logs a message alongsize a serialized group.
func groupLogf(g *zipkey.Group, s string, vs ...interface{}) {
	log.Printf(s, vs...)
	b, _ := json.Marshal(g)
	log.Println(string(b))
}

// ZippyExact takes a release and refs reader (key, doc) and assigns a fixed
// match result.
func ZippyExact(releases, refs io.Reader, matchResult MatchResult, w io.Writer) error {
	var (
		enc     = json.NewEncoder(w)
		keyer   = makeKeyFunc("\t", 1)
		i       = 0
		grouper = func(g *zipkey.Group) error {
			i++
			if i%10000 == 0 {
				log.Printf("processed %v groups", i)
			}
			var (
				target *Release
				ref    *Ref
				err    error
			)
			if len(g.G0) == 0 || len(g.G1) == 0 {
				return nil
			}
			if target, err = stringToRelease(Cut(g.G0[0], 2)); err != nil {
				groupLogf(g, "[skip] failed to parse release: %v", err)
				return nil
			}
			for _, line := range g.G1 {
				if ref, err = stringToRef(Cut(line, 2)); err != nil {
					groupLogf(g, "[skip] failed to parse ref: %v", err)
					continue
				}
				var bref BiblioRef
				bref.SourceReleaseIdent = ref.ReleaseIdent
				bref.SourceWorkIdent = ref.WorkIdent
				bref.SourceReleaseStage = ref.ReleaseStage
				bref.SourceYear = fmt.Sprintf("%d", ref.ReleaseYear)
				bref.RefIndex = ref.Index + 1 // we want 1-index (also helps with omitempty)
				bref.RefKey = ref.Key
				bref.TargetReleaseIdent = target.Ident
				bref.TargetWorkIdent = target.WorkID
				bref.MatchProvenance = ref.RefSource
				bref.MatchStatus = matchResult.Status.Short()
				bref.MatchReason = matchResult.Reason.Short()
				if err := enc.Encode(bref); err != nil {
					return err
				}
			}
			return nil
		}
	)
	zipper := zipkey.New(releases, refs, keyer, grouper)
	return zipper.Run()
}

// ZippyExactWiki takes a release and wiki reader (key, doc) and assigns a
// fixed match result.
func ZippyExactWiki(releases, wiki io.Reader, mr MatchResult, w io.Writer) error {
	var (
		enc     = json.NewEncoder(w)
		keyer   = makeKeyFunc("\t", 1)
		grouper = func(g *zipkey.Group) error {
			var (
				target *Release
				wiki   *MinimalCitations
				err    error
			)
			if len(g.G0) == 0 || len(g.G1) == 0 {
				return nil
			}
			if target, err = stringToRelease(Cut(g.G0[0], 2)); err != nil {
				return err
			}
			for _, line := range g.G1 {
				if wiki, err = stringToWiki(Cut(line, 2)); err != nil {
					return err
				}
				var bref BiblioRef
				bref.Key = fmt.Sprintf("%s_%s", slugifyString(wiki.PageTitle), target.Ident) // XXX: what should we use?
				bref.SourceWikipediaArticle = wiki.PageTitle
				bref.TargetReleaseIdent = target.Ident
				bref.TargetWorkIdent = target.WorkID
				bref.MatchProvenance = "wikipedia"
				bref.MatchStatus = mr.Status.Short()
				bref.MatchReason = mr.Reason.Short()
				if err := enc.Encode(bref); err != nil {
					return err
				}
			}
			return nil
		}
	)
	zipper := zipkey.New(releases, wiki, keyer, grouper)
	return zipper.Run()
}

// ZippyVerifyRefs takes a release and refs reader (key, doc), run fuzzy
// verification and will emit a biblioref document, if exact or strong match.
func ZippyVerifyRefs(releases, refs io.Reader, w io.Writer) error {
	var (
		enc     = json.NewEncoder(w)
		keyer   = makeKeyFunc("\t", 1)
		grouper = func(g *zipkey.Group) error {
			var (
				re, pivot *Release
				err       error
			)
			if len(g.G0) == 0 || len(g.G1) == 0 {
				return nil
			}
			if pivot, err = stringToRelease(Cut(g.G0[0], 2)); err != nil {
				return err
			}
			for _, line := range g.G1 {
				if re, err = stringToRelease(Cut(line, 2)); err != nil {
					return err
				}
				result := Verify(pivot, re)
				switch result.Status {
				case StatusExact, StatusStrong:
					if result.Reason == ReasonDOI {
						continue
					}
					br := generateBiblioRef(re, pivot, result, "fuzzy")
					if err := enc.Encode(br); err != nil {
						return err
					}
				default:
					// XXX: We want to add unmatched pieces as well; here?  We
					// probably want to do a single final pass to complete the
					// dataset.
				}
			}
			return nil
		}
	)
	zipper := zipkey.New(releases, refs, keyer, grouper)
	return zipper.Run()
}

// ZippyVerifyRefsOpenLibraryTable takes OL editions (as release) and refs (as
// release) and emits a match table for manual inspection.
func ZippyVerifyRefsOpenLibraryTable(olr, refs io.Reader, w io.Writer) error {
	var (
		keyer   = makeKeyFunc("\t", 1)
		grouper = func(g *zipkey.Group) error {
			var (
				re, pivot *Release
				err       error
			)
			if len(g.G0) == 0 || len(g.G1) == 0 {
				return nil
			}
			// We take a single edition from OL.
			if pivot, err = stringToRelease(Cut(g.G0[0], 2)); err != nil {
				return err
			}
			for _, line := range g.G1 {
				if re, err = stringToRelease(Cut(line, 2)); err != nil {
					return err
				}
				// The refs have a container name, but not a title, but here we
				// compare against titles from open library.
				re.Title = re.ContainerName
				result := Verify(pivot, re)
				fmt.Printf("%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\n",
					result.Status.Short(),
					result.Reason.Short(),
					pivot.Extra.OpenLibrary.WorkID,
					FindByPrefix(pivot.Extra.OpenLibrary.SourceRecords, "ia:"),
					re.Ident,
					CutSep(g.G0[0], "\t", 1),
					pivot.Title,
					re.Title)
			}
			return nil
		}
	)
	zipper := zipkey.New(olr, refs, keyer, grouper)
	return zipper.Run()
}

// ZippyVerifyRefsOpenLibrary takes OL editions (as release) and refs (as
// release) and emits a match table for manual inspection.
func ZippyVerifyRefsOpenLibrary(olr, refs io.Reader, w io.Writer) error {
	var (
		enc     = json.NewEncoder(w)
		keyer   = makeKeyFunc("\t", 1)
		grouper = func(g *zipkey.Group) error {
			var (
				ref, pivot *Release // ref (reference), pivot (open library)
				err        error
			)
			if len(g.G0) == 0 || len(g.G1) == 0 {
				return nil
			}
			// We take a single edition from OL.
			if pivot, err = stringToRelease(Cut(g.G0[0], 2)); err != nil {
				return err
			}
			for _, line := range g.G1 {
				if ref, err = stringToRelease(Cut(line, 2)); err != nil {
					return err
				}
				// The refs have a container name, but not a title, but here we
				// compare against titles from open library.
				ref.Title = ref.ContainerName
				result := Verify(pivot, ref)
				switch result.Status {
				case StatusExact, StatusStrong:
					var bref BiblioRef
					bref.SourceReleaseIdent = ref.Ident
					bref.SourceWorkIdent = ref.WorkID
					bref.SourceReleaseStage = ref.ReleaseStage
					bref.SourceYear = fmt.Sprintf("%d", ref.ReleaseYear())
					bref.RefIndex = ref.Extra.Skate.Ref.Index + 1 // we want 1-index (also helps with omitempty)
					bref.RefKey = ref.Extra.Skate.Ref.Key
					bref.TargetOpenLibraryWork = pivot.WorkID
					bref.MatchProvenance = ref.Extra.Skate.Ref.Source
					bref.MatchStatus = result.Status.Short()
					bref.MatchReason = result.Reason.Short()
					if err := enc.Encode(bref); err != nil {
						return err
					}
				default:
				}
			}
			return nil
		}
	)
	zipper := zipkey.New(olr, refs, keyer, grouper)
	return zipper.Run()
}

// Cut returns a specific column (1-indexed, like CutSep) from a tabular
// file, returns empty string if column is invalid.
func Cut(line string, column int) string {
	return CutSep(line, "\t", column)
}

// CutSep allows to specify a separator.
func CutSep(line, sep string, column int) string {
	parts := strings.Split(strings.TrimSpace(line), sep)
	if len(parts) < column {
		return ""
	} else {
		return parts[column-1]
	}
}

// FindByPrefix return the first element for a slice of strings, which matches a prefix.
func FindByPrefix(ss []string, prefix string) string {
	for _, s := range ss {
		if strings.HasPrefix(s, prefix) {
			return s
		}
	}
	return ""
}

// makeKeyFunc creates a function that can be used as keyFunc, selecting a
// column from fields separated by sep; column is 1-indexed.
func makeKeyFunc(sep string, column int) func(string) (string, error) {
	return func(s string) (string, error) {
		if k := CutSep(s, sep, column); k == "" {
			return k, fmt.Errorf("cannot get key from column %d in line (len=%d): %s", column, len(s), s)
		} else {
			return k, nil
		}
	}
}

func stringToRelease(s string) (r *Release, err error) {
	err = json.Unmarshal([]byte(s), &r)
	return
}

func stringToRef(s string) (r *Ref, err error) {
	err = json.Unmarshal([]byte(s), &r)
	return
}

func stringToWiki(s string) (r *MinimalCitations, err error) {
	err = json.Unmarshal([]byte(s), &r)
	return
}