The core process performs the primary scoring of edits using an artificial neural network.
All request and response packets are based on XML and follow a fixed schema; all fields are expected with relevant values.
<?xml version="1.0"?>
<WPEditSet>
<WPEdit>
<EditType>change</EditType>
<EditID>1022572696</EditID>
<comment>/* Political career */</comment>
<user>JamesVilla44</user>
<user_edit_count>9146</user_edit_count>
<user_distinct_pages />
<user_warns>1</user_warns>
<prev_user>JamesVilla44</prev_user>
<user_reg_time>1553976920</user_reg_time>
<common>
<page_made_time>1620437208</page_made_time>
<title>Angelique Foster</title>
<namespace>Main:</namespace>
<creator>Moondragon21</creator>
<num_recent_edits>14</num_recent_edits>
<num_recent_reversions>0</num_recent_reversions>
</common>
<current>
<minor>false</minor>
<timestamp>1620720638</timestamp>
<text>The current contents</text>
</current>
<previous>
<timestamp>1620720514</timestamp>
<text>The previous contents</text>
</previous>
</WPEdit>
</WPEditSet>
<WPEditSet>
<WPEdit>
<editid>1022572696</editid>
<score>0.149043</score>
<think_vandalism>false</think_vandalism>
</WPEdit>
</WPEditSet>
There are 2 primary artifacts required for the process to execute;
- Configuration (
conf/
) - Trained data set (
data/
)
Both sets will be compiled as required & included in the generated container/build artifacts.
To provide a consistent environment, the current build logic is contained within a docker container.
Local complication can be performed via docker build .
, the CI workflow will also extract the individual binaries for direct consumption.