Introduce textblob framework to analysis word

[mrproliu]

If the pattern contains words, then we should not combine them as a variable path.
For example, there are three URLs /test/t/one, test/t/two/, and test/t/three. In the old way, they will combine to /test/t/{var}, After we add textblob and word spitter, the result will keep the original URL list.

The Basic logic is:

1. Split the word in each URL step by camel and digits.
2. Use the textblob framework to analyze whether the word is correct or not, If the word is correct, then the URL will not merge into the existing pattern, otherwise, keep the original logical.

Finally, we can reduce the DRAIN_COMBINE_MIN_URL_COUNT environment to reduce the count of combined URL logic.

[wu-sheng]

Could you share the examples here? Such as how these work

• /rootpath/productSale
• /rootpath/prodcut_sale/
• /rootpath/product/{UUID list}, UUID is a combination of digits and texts, but not meaningful words.

[mrproliu]

Here are some examples:

All word

Data

/test/productSale
/test/product_Sale
/test/productOnSale

Result

/test/productSale
/test/product_Sale
/test/productOnSale

UUID

Data

/test/one/f0dce0206a8011ef8670df368ea4fcf1
/test/one/f0dce0216a8011ef8670df368ea4fcf1
/test/one/f0dce0226a8011ef8670df368ea4fcf1
/test/two/0730f8706a8111ef8670df368ea4fcf1
/test/two/0730f8716a8111ef8670df368ea4fcf1
/test/two/0730f8726a8111ef8670df368ea4fcf1
/test/three/26e2add06a8111ef8670df368ea4fcf1
/test/three/26e2d4e06a8111ef8670df368ea4fcf1
/test/three/26e2d4e16a8111ef8670df368ea4fcf1

Result

/test/one/{var}
/test/two/{var}
/test/three/{var}