JetBrains.Platform.Util.Text
intentionally named in wrong order :)
Combine score and adjustment with coefficient.
Negates sum to obtain comparable int.
Use positive/negative coefficient to increase/decrease rank of item in search results
Text scoring honor Exact middle match more, than match by humps
Matches text by pattern using so called CamelHump. Not immutable, has internal state.
!!! MUST BE used in single-threaded fashion, because every processing method changes internal state.
Caution!!! Use it only for stream processing to reduce memory traffic or with defensive copying.
This is shared stated and it will be changed after next invocation.
Matches with minimal required flags in
declare property of matcher, that if wv is matched, where w and v are words, then w is also matched
Used for caching purposes
Implements Boyer-Moore search algorithm (case-sensitive)
This method calculate for each pattern character the list of possible indices of input string.
The resulting graph.
For example, for input string "pj ppe" and pattern ppe it looks like as follow
Note: there is not zeroing memory of all graph for performance reasons. g means that in that part of array garbage is located.
-1 means end of list.
| input index | 0 | 1 | 2 | 3 | 4 | 5 |
| input char | p | j | | p | p | e |
| ------------- | --- | --- | --- | --- | --- | --- |
| p | 0 | 3 | -1 | g | g | g |
| p | 0 | 3 | 4 | -1 | g | g |
| e | 5 | -1 | g | g | g | g |
length of input
the pointer to lowered input text
the unsafe array of humps indices.
the pointer to original input text
if corresponding character in input string is found for each character in pattern.
This method calculates the best matching indices of input string to pattern using dynamic programming when there is guarantee that
a match with correct indices exist.
the matrix of possible corresponding chars of input string to pattern (see detailed description above)
input length
mt is resulting array of input length
hump lookup for input string
This is externally provided storage for dynamics.
This array threaded as matrix with row count = pattern length, column count = input length.
For example, the following matrix is calculated for pattern ppe and input string "pj ppe"
h = humpScore = 10
s = nonHumpScore = 1
c = letter count score = 100000 - fake scoring for ensure that every symbol of input is corresponding to character of input text
path direction is determined by the sign of score:
Negative = maximal score taken from skipping this character of input string.
Positive = maximal score taken from matching this character of input string.
| input index | 0 | 1 | 2 | 3 | 4 | 5 |
| input char | p | j | | p | p | e |
| ------------- | --- | --- | --- | --- | --- | --- |
| p | -c-h | c+h | c+h | -c-h | c+h | c+h |
| p | 0 | 0 | 0 | -2c-2h | -2c-h-s | 2c+h+s |
| e | 0 | 0 | 0 | 0 | 0 | 3c+1h+2s |
Create wildcard identifier mather. Can be used together with * and ? to create exact matches
the wildcard pattern for
Unescaped pattern value with quotes to ability create correct clones with Identifier matchers factories.
Creates short abbreviated names for the entities.
Treats just string length.
Produces abbreviated names for the given set of items.
The list of items to process.
The way of getting the item names.
Max number of characters allowed in the output.
Tells whether the char is CAPS, in our terms.
Initial data.
Just limits all items to the given maxlen (at most).
Take chars at most (this includes as many caps as possible).
Gets the item name, splits it into parts by the Caps.
Each part is a property bag.
Reverses the internal map to make it item-to-name.
When shortening the abbreviations, we've taken the first chars from each capsed part.
Try taking some non-first chars instead, if they help making the abbr unique.
Appends disambig chars to the end of the abbreviation, as a last resort.