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
|
local log = require('telescope.log')
local util = require('telescope.utils')
local sorters = {}
local ngram_highlighter = function(ngram_len, prompt, display)
local highlights = {}
display = display:lower()
for disp_index = 1, #display do
local char = display:sub(disp_index, disp_index + ngram_len - 1)
if prompt:find(char, 1, true) then
table.insert(highlights, {
start = disp_index,
finish = disp_index + ngram_len - 1
})
end
end
return highlights
end
local Sorter = {}
Sorter.__index = Sorter
---@class Sorter
--- Sorter sorts a list of results by return a single integer for a line,
--- given a prompt
---
--- Lower number is better (because it's like a closer match)
--- But, any number below 0 means you want that line filtered out.
--- @field scoring_function function Function that has the interface:
-- (sorter, prompt, line): number
function Sorter:new(opts)
opts = opts or {}
return setmetatable({
state = {},
scoring_function = opts.scoring_function,
highlighter = opts.highlighter,
}, Sorter)
end
function Sorter:score(prompt, entry)
if not entry or not entry.ordinal then return -1 end
return self:scoring_function(prompt or "", entry.ordinal, entry)
end
function sorters.new(...)
return Sorter:new(...)
end
sorters.Sorter = Sorter
TelescopeCachedTails = TelescopeCachedTails or nil
if not TelescopeCachedTails then
local os_sep = util.get_separator()
local match_string = '[^' .. os_sep .. ']*$'
TelescopeCachedTails = setmetatable({}, {
__index = function(t, k)
local tail = string.match(k, match_string)
rawset(t, k, tail)
return tail
end,
})
end
TelescopeCachedUppers = TelescopeCachedUppers or setmetatable({}, {
__index = function(t, k)
local obj = {}
for i = 1, #k do
local s_byte = k:byte(i, i)
if s_byte <= 90 and s_byte >= 65 then
obj[s_byte] = true
end
end
rawset(t, k, obj)
return obj
end
})
TelescopeCachedNgrams = TelescopeCachedNgrams or {}
-- TODO: Match on upper case words
-- TODO: Match on last match
sorters.get_fuzzy_file = function(opts)
opts = opts or {}
local ngram_len = opts.ngram_len or 2
local function overlapping_ngrams(s, n)
if TelescopeCachedNgrams[s] and TelescopeCachedNgrams[s][n] then
return TelescopeCachedNgrams[s][n]
end
local R = {}
for i = 1, s:len() - n + 1 do
R[#R+1] = s:sub(i, i+n-1)
end
if not TelescopeCachedNgrams[s] then
TelescopeCachedNgrams[s] = {}
end
TelescopeCachedNgrams[s][n] = R
return R
end
return Sorter:new {
scoring_function = function(_, prompt, line)
local N = #prompt
if N == 0 or N < ngram_len then
-- TODO: If the character is in the line,
-- then it should get a point or somethin.
return 0
end
local prompt_lower = prompt:lower()
local line_lower = line:lower()
local prompt_lower_ngrams = overlapping_ngrams(prompt_lower, ngram_len)
-- Contains the original string
local contains_string = line_lower:find(prompt_lower, 1, true)
local prompt_uppers = TelescopeCachedUppers[prompt]
local line_uppers = TelescopeCachedUppers[line]
local uppers_matching = 0
for k, _ in pairs(prompt_uppers) do
if line_uppers[k] then
uppers_matching = uppers_matching + 1
end
end
-- TODO: Consider case senstivity
local tail = TelescopeCachedTails[line_lower]
local contains_tail = tail:find(prompt, 1, true)
local consecutive_matches = 0
local previous_match_index = 0
local match_count = 0
for i = 1, #prompt_lower_ngrams do
local match_start = line_lower:find(prompt_lower_ngrams[i], 1, true)
if match_start then
match_count = match_count + 1
if match_start > previous_match_index then
consecutive_matches = consecutive_matches + 1
end
previous_match_index = match_start
end
end
local tail_modifier = 1
if contains_tail then
tail_modifier = 2
end
local denominator = (
(10 * match_count / #prompt_lower_ngrams)
-- biases for shorter strings
+ 3 * match_count * ngram_len / #line
+ consecutive_matches
+ N / (contains_string or (2 * #line))
-- + 30/(c1 or 2*N)
-- TODO: It might be possible that this too strongly correlates,
-- but it's unlikely for people to type capital letters without actually
-- wanting to do something with a capital letter in it.
+ uppers_matching
) * tail_modifier
if denominator == 0 or denominator ~= denominator then
return -1
end
if #prompt > 2 and denominator < 0.5 then
return -1
end
return 1 / denominator
end,
highlighter = opts.highlighter or function(_, prompt, display)
return ngram_highlighter(ngram_len, prompt, display)
end,
}
end
sorters.get_generic_fuzzy_sorter = function(opts)
opts = opts or {}
local ngram_len = 2
local function overlapping_ngrams(s, n)
if TelescopeCachedNgrams[s] and TelescopeCachedNgrams[s][n] then
return TelescopeCachedNgrams[s][n]
end
local R = {}
for i = 1, s:len() - n + 1 do
R[#R+1] = s:sub(i, i+n-1)
end
if not TelescopeCachedNgrams[s] then
TelescopeCachedNgrams[s] = {}
end
TelescopeCachedNgrams[s][n] = R
return R
end
return Sorter:new {
-- self
-- prompt (which is the text on the line)
-- line (entry.ordinal)
-- entry (the whole entry)
scoring_function = function(_, prompt, line, _)
if prompt == 0 or #prompt < ngram_len then
return 0
end
local prompt_lower = prompt:lower()
local line_lower = line:lower()
local prompt_ngrams = overlapping_ngrams(prompt_lower, ngram_len)
local N = #prompt
local contains_string = line_lower:find(prompt_lower, 1, true)
local consecutive_matches = 0
local previous_match_index = 0
local match_count = 0
for i = 1, #prompt_ngrams do
local match_start = line_lower:find(prompt_ngrams[i], 1, true)
if match_start then
match_count = match_count + 1
if match_start > previous_match_index then
consecutive_matches = consecutive_matches + 1
end
previous_match_index = match_start
end
end
-- TODO: Copied from ashkan.
local denominator = (
(10 * match_count / #prompt_ngrams)
-- biases for shorter strings
-- TODO(ashkan): this can bias towards repeated finds of the same
-- subpattern with overlapping_ngrams
+ 3 * match_count * ngram_len / #line
+ consecutive_matches
+ N / (contains_string or (2 * #line))
-- + 30/(c1 or 2*N)
)
if denominator == 0 or denominator ~= denominator then
return -1
end
if #prompt > 2 and denominator < 0.5 then
return -1
end
return 1 / denominator
end,
highlighter = opts.highlighter or function(_, prompt, display)
return ngram_highlighter(ngram_len, prompt, display)
end,
}
end
-- Bad & Dumb Sorter
sorters.get_levenshtein_sorter = function()
return Sorter:new {
scoring_function = function(_, prompt, line)
return require('telescope.algos.string_distance')(prompt, line)
end
}
end
return sorters
|