|
| 1 | +# FuzzyMap <img src="https://avatars.githubusercontent.com/u/108220139" align="right" height="64" /> |
| 2 | + |
| 3 | +[](https://pypi.org/project/fuzzymap/) |
| 4 | +[](https://github.com/pysnippet/fuzzymap/blob/master/LICENSE) |
| 5 | + |
| 6 | +## What is FuzzyMap? |
| 7 | + |
| 8 | +`FuzzyMap` is a polymorph Python dictionary. This kind of dictionary returns the value of the exact key if there is such |
| 9 | +a key. Otherwise, it will return the value of the most similar key satisfying the given ratio. The same mechanism works |
| 10 | +when setting a new or replacing an old key in the dictionary. If the key is not found and does not match any of the keys |
| 11 | +by the given ratio, it returns `None`. |
| 12 | + |
| 13 | +## How does it work? |
| 14 | + |
| 15 | +Suppose you have scraped data from multiple sources that do not have a unique identifier, and you want to compare the |
| 16 | +values of the items having the same identifiers. Sure there will be found a field that mostly has an equivalent value |
| 17 | +at each source. And you can use that field to identify the corresponding items of other sources' data. |
| 18 | + |
| 19 | +## Let's look at the following example |
| 20 | + |
| 21 | +There is a live data parser that collects the coefficients of football matches from different bookmakers at once, then |
| 22 | +calculates and logs the existing forks. Many bookmakers change the name of the teams to be incomparable with names on |
| 23 | +other sites. |
| 24 | + |
| 25 | +```python |
| 26 | +from fuzzymap import FuzzyMap |
| 27 | + |
| 28 | +src1 = { |
| 29 | + 'Rapid Wien - First Vienna': {'w1': 1.93, 'x': 2.32, 'w2': 7.44}, |
| 30 | + 'Al Bourj - Al Nejmeh': {'w1': 26, 'x': 11.5, 'w2': 1.05}, |
| 31 | + # hundreds of other teams' data |
| 32 | +} |
| 33 | + |
| 34 | +src2 = FuzzyMap({ |
| 35 | + 'Bourj FC - Nejmeh SC Beirut': {'w1': 32, 'x': 12, 'w2': 1.05}, |
| 36 | + 'SK Rapid Wien - First Vienna FC': {'w1': 1.97, 'x': 2.3, 'w2': 8.2}, |
| 37 | + # hundreds of other teams' data |
| 38 | +}) |
| 39 | + |
| 40 | +for team, coefs1 in src1.items(): |
| 41 | + coefs2 = src2[team] |
| 42 | + |
| 43 | + # coefs1 = {"w1": 1.93, "x": 2.32, "w2": 7.44} |
| 44 | + # coefs2 = {"w1": 1.97, "x": 2.3, "w2": 8.2} |
| 45 | + handle_fork(coefs1, coefs2) |
| 46 | +``` |
| 47 | + |
| 48 | +With a human brain, it is not difficult to identify that "Rapid Wien - First Vienna" and "SK Rapid Wien - First Vienna |
| 49 | +FC" matches are the same. In the above example, the `src2` is defined as `FuzzyMap`, it makes its keys fuzzy-matchable, |
| 50 | +and we can get an item corresponding to the key of `src1`. See the below graph demonstrating the associations of |
| 51 | +`FuzzyMap` keys. |
| 52 | + |
| 53 | +```mermaid |
| 54 | +graph LR |
| 55 | + src1team1[Rapid Wien - First Vienna]-->src1coefs1["{'w1': 1.93, 'x': 2.32, 'w2': 7.44}"] |
| 56 | + src1team2[Al Bourj - Al Nejmeh]-->src1coefs2["{'w1': 26, 'x': 11.5, 'w2': 1.05}"] |
| 57 | + src2team1[SK Rapid Wien - First Vienna FC]-->src2coefs1["{'w1': 1.97, 'x': 2.3, 'w2': 8.2}"] |
| 58 | + src2team2[Bourj FC - Nejmeh SC Beirut]-->src2coefs2["{'w1': 32, 'x': 12, 'w2': 1.05}"] |
| 59 | + src1team1-->src2coefs1 |
| 60 | + src1team2-->src2coefs2 |
| 61 | +``` |
| 62 | + |
| 63 | +## License |
| 64 | + |
| 65 | +Copyright (C) 2022 Artyom Vancyan. [GPLv2](LICENSE) |
0 commit comments