vim_sentence_auto_complete

A sentence auto completer plugin for VIM with a file containing most frequent sentences. A C++ utility is used to get the most frequent sentences from a given data set.

Plugin was developed to ease up writing repeated JAMA automated test descriptions.

Most frequent sentences are collected using trie and heap data structure for faster performance and low memory foot print for large data sets.

Generating most frequent sentences from a dataset/github repo

1. Compile automater.cpp

g++ automater.cpp -std=c++17 -lstdc++fs 

2. Get the most frequent sentences from an existing folder train_data or freshly clone a repo from github you wish to collect most frequent lines from(set your appropriate github username and authtoken in the source file). Training data file pattern, description start/stop are hardcoded in the file which can be modified.
   

Sample command:
use existing train data without cloning any repo and get the top 100 frequent sentences.

./a.out none sanitizePatterns.txt 0 1 100

3. The output will be a file mostFeqLines.txt containing tab separated sentences with frequency.

Using the plugin

1. Filter the output from previous steps as per your requirement and create a file most_freq_lines.txt with the final list of sentences that will be used for VIM autocomplete.

2. Make sure the plugin file .vim and most_freq_lines.txt are in the same directory. You can source it in .vimrc or source it using :source hari_autocomplete.vim from your current buffer.

3. Once sourced use :SSC to active the new omni function and :USSC to reset it. Completion can be triggered using Ctrl-x Ctrl-o in insert mode. Both the lines from current file and most_freq_lines.txt are used in auto completion.