Rule-based-Address-Parser

An Vietnamese address parser that based on rules.

Warning This repo is not maintained. Please visit the issues tab for data related questions.

Example
Requirements
Running & How it works
Acknowledgments

Example

It can parse a Vietnamese address into standard one.

Input:

107 Bạch Đằng, Âu Cơ, Phu Tho, 35000, Vietnam

Ouput:

"house_number": "107",
"house_number_type": "Số",
"post_code": "35000",
"country": "Việt Nam",
"street": "Bạch Đằng",
"street_type": "Đường",
"ward": "Âu Cơ",
"ward_type": "Phường",
"town": "Phú Thọ",
"town_type": "Thị xã",
"province": "Phú Thọ",
"province_type": "Tỉnh"

Requirements

Python 3.7.7
xlrd

Running & How it works

Export the standard addresses (from province level to ward level)

>> python excel_parser/excel_parser.py

Export the standard address tree (The address tree has country as the root node and each child node of a node has a smaller address level)

In addr_graph.py, uncomment this line and comment the rest:

std_addr_graph = export_std_addr_graph()

Then, execute the following command:

>> python address_graph/addr_graph.py

Clean the raw address data

>> python ultis/clean_data.py

Parse the cleaned address data

>> python json_parser/json_parser.py

Lastly, use standard address tree to standardize the address (export to a Reference Address Table - RAT):

In addr_graph.py, uncomment this lines and comment the rest:

std_addr_graph.export_to_RAT(p._CLEAN_DATA_DIR_PATH + p.RAT_JSON_PATH)