<h1>Pipeline Algorithm</h1>
<ul>
<li>
<h3>Basic CRUD Operations</h3>
<ol>
<li>
<h4>Create</h4>
<ol>
<li>
<b>The frontend</b>
<ul>
<li>Opens the Prompt Registry pane</li>
<li>Validates the payload before sending it to the server</li>
<li>Sends a POST request to the backend</li>
</ul>
</li>
<li>
<b>The backend</b>
<ul>
<li>receives the request</li>
<li>Validates the payload</li>
<li>Stores the record in the database</li>
<li>Sends a response to the frontend</li>
</ul>
</li>
<li>
<b>The frontend</b>
<ul>
<li>Receives the response from the backend</li>
<li>Notifies the user of succes or failure</li>
</ul>
</li>
</ol>
</li>
<li>
<h4>Read</h4>
<ol>
<li>
<b>The frontend</b>
<ul>
<li>Opens the Prompt Registry pane</li>
<li>Sends a GET request to the backend, requesting data</li>
</ul>
</li>
<li>
<b>The backend</b>
<ul>
<li>Receives the request</li>
<li>Validates the payload</li>
<li>Fetches the record from the database</li>
<li>Sends a response to the frontend</li>
</ul>
</li>
<li>
<b>The frontend</b>
<ul>
<li>Receives the response from the backend</li>
<li>Hydrates the Promp Registry</li>
</ul>
</li>
</ol>
</li>
<li>
<h4>Update</h4>
<ol>
<li>
<b>The frontend</b>
<ul>
<li>Opens the Prompt Registry pane</li>
<li>Makes local changes to an existing record</li>
<li>Validates the payload before sending it to the server</li>
<li>Sends an UPDATE request to the backend</li>
</ul>
</li>
<li>
<b>The backend</b>
<ul>
<li>Receives the request</li>
<li>Validates the payload</li>
<li>Updates the record in the database</li>
<li>Sends a response to the frontend</li>
</ul>
</li>
<li>
<b>The frontend</b>
<ul>
<li>Receives the response from the backend</li>
<li>Notifies the user of success or failure</li>
</ul>
</li>
</ol>
</li>
<li>
<h4>Delete</h4>
<ol>
<li>
<b>The frontend</b>
<ul>
<li>Opens the Prompt Registry pane</li>
<li>Locates an existing record</li>
<li>Validates the payload before sending it to the server</li>
<li>Sends a DELETE request to the backend</li>
</ul>
</li>
<li>
<b>The backend</b>
<ul>
<li>Receives the request</li>
<li>Validates the payload</li>
<li>Deletes the record from the database</li>
<li>Sends a response to the frontend</li>
</ul>
</li>
<li>
<b>The frontend</b>
<ul>
<li>Receives the response from the backend</li>
<li>Notifies the user of success or failure</li>
</ul>
</li>
</ol>
</li>
</ol>
</li>
<li>
<h3>Compiler Job</h3>
<ol>
<li>
<h4>Frontend Initialization</h4>
<ul>
<li>
Opens the compiler configuration pane and establishes settings.
</li>
<li>Opens the terminal and selects the target script.</li>
<li>Validates the payload data structure.</li>
</ul>
</li>
<li>
<h4>Job Dispatch</h4>
<ul>
<li>Frontend transmits the payload to the server via HTTP POST.</li>
<li>Server receives and validates the incoming payload.</li>
<li>
Server responds with a unique Job ID and begins asynchronous
execution.
</li>
<li>
Frontend subscribes to the task by initiating a polling loop (HTTP
GET) using the Job ID.
</li>
</ul>
</li>
<li>
<h4>Native Compilation</h4>
<ul>
<li>
Server passes the validated payload to the C++ compiler via
Pybind11.
</li>
<li>
Compiler executes the pipeline and saves output artifacts to disk.
</li>
<li>
Compiler returns the execution manifest back across the Pybind11
boundary.
</li>
</ul>
</li>
<li>
<h4>State Synchronization</h4>
<ul>
<li>
Backend receives the manifest and persists it to the database.
</li>
<li>
Frontend polling loop detects the job status transition from
"building" to "success".
</li>
<li>
Frontend issues a final GET request to fetch the manifest and
hydrate the interface.
</li>
</ul>
</li>
</ol>
</li>
<li>
<h3>Neural Request</h3>
<ol>
<li>
<h4>Frontend Initialization</h4>
<ul>
<li>Opens the prompt dashboard.</li>
<li>Selects a designated PromptPair and target Python script.</li>
<li>Validates the payload data structure.</li>
</ul>
</li>
<li>
<h4>Backend Dispatch</h4>
<ul>
<li>Frontend transmits the payload to the server via HTTP POST.</li>
<li>Backend receives and validates the incoming request.</li>
<li>Backend persists the initial payload state in the database.</li>
</ul>
</li>
<li>
<h4>LLM Execution & Persistence</h4>
<ul>
<li>
Backend routes the payload to the designated local LLM server via
HTTP POST.
</li>
<li>
LLM server processes the prompt and returns the generated inference.
</li>
<li>
Backend receives the inference result and updates the database
record.
</li>
</ul>
</li>
<li>
<h4>State Synchronization</h4>
<ul>
<li>Backend returns the final inference response to the frontend.</li>
<li>
Frontend renders the comparative results within the dual-pane
Inference Window.
</li>
</ul>
</li>
</ol>
</li>
</ul>