Skip to content

2.3.7. Mission Control (The GalaxyScope Orchestrator)

Phase 0: Mission Control

The GalaxyScope Orchestrator (galaxyscope.py) is the central nervous system of the blAST engine. It acts as Mission Control, managing the flow of data from the initial file-system radar ping down to the final serialization of the 3D map. Its primary responsibility is maximizing computational velocity while preventing catastrophic memory leaks or regex deadlocks during hyper-scale scans.

2.3.7.A. Multi-Core Extraction (The Worker Pool)

Parsing millions of lines of code on a single thread is a bottleneck. To achieve hyper-scale velocity (100,000+ LOC/s), the Orchestrator implements a highly tuned ProcessPoolExecutor Map-Reduce architecture.

  • Isolated Memory Spaces: The _init_worker function boots copies of the heavy regex engines (LogicSplicer, Prism, LanguageDetector) directly into the isolated memory of each CPU core. This prevents cross-thread lockups.
  • Cache Warming: The workers force-warm the regex compilers for the specific languages found in the repository during the radar ping, completely eliminating the "Plaintext Stutter" (lazy-loading lag) on the first few thousand files.
  • The Phantom Check: Workers perform an instantaneous disk-check before parsing. If a file was reported by Git but has since vanished from the disk, it is silently evaporated as a "Phantom," preventing the main thread from logging false anomalies.

2.3.7.B. The Starvation Monitor (ReDoS Shield)

When scanning massive, auto-generated codebases, poorly written files can occasionally trigger Catastrophic Regex Backtracking (ReDoS), which locks a CPU core in an infinite loop.

  • The 60-Second Guillotine: Instead of blindly waiting for the worker pool to finish, the Orchestrator uses a concurrent.futures.wait loop. If all CPU workers freeze and fail to return a single processed file within 60 seconds, the Starvation Monitor detects the deadlock.
  • Graceful Abort: It instantly logs the exact file paths that caused the hang, relegates them to the Singularity audit log, and forcefully shuts down the executor pool to unfreeze the user's terminal.

2.3.7.C. \(O(1)\) Relational Resolution (Pass 1.5)

To build the dependency graph, the engine must figure out which files are importing which other files. Cross-referencing thousands of raw import strings (e.g., import utils) against thousands of file paths (e.g., src/core/utils.py) originally created an exponential \(O(N^2)\) compute bomb.

  • The Suffix Hash Map: Pass 1.5 defuses this bomb. It pre-computes every possible valid path suffix for every file in the repository and stores them in a lightning-fast hash map.
  • \(O(1)\) Lookups: When evaluating a raw import string, the engine simply checks if the string exists as a key in the map. This turns an exponential search into an instantaneous \(O(1)\) lookup, allowing GitGalaxy to resolve millions of dependency links in milliseconds.

2.3.7.D. Relational Context & Mass Dampeners (Pass 2)

In Pass 2, the Orchestrator evaluates files based on their surrounding neighborhood before handing them off to the Signal Processor.

  • Domain Ontologies: It tallies the languages in every folder to determine the "Dominant Ecosystem," allowing downstream engines to spot Trojan files (e.g., a C++ file hiding in a JavaScript folder).
  • The Umbrella Shield: It calculates the total percentage of the repository dedicated to testing (/tests/ folders). If a repo is heavily tested globally, the Orchestrator applies an "Umbrella Bonus" that mathematically dampens the testing risk for individual logic files.
  • Structural Mass Dampeners: It intercepts files acting as heavy, non-executable data dumps (like Wycheproof test vectors, vendored Kubernetes code-gen, or massive translation dictionaries) and physically crushes their mass multiplier so they do not mathematically eclipse the human-written architecture in the 3D view.

2.3.7.E. The CLI & Smart Threat Switch

The main() entry point manages user interaction and environment overrides.

  • Dialect Injection: It intercepts the PROJECT_OVERRIDES registry. If the user is scanning a known legacy repository (like the FreeBSD kernel), it dynamically hot-patches the LANGUAGE_DEFINITIONS in RAM to parse its specific dialect perfectly.
  • The Smart Threat Switch (--paranoid): If invoked, the Orchestrator loads the "Hazmat Suit" threat policy, lowering the threshold for the Security Lens to flag deeply hidden steganography and logic bombs.
  • Instant RAM Eviction: Once the artifacts are sealed to the disk, the CLI invokes a hard os._exit(0). This violently drops the Python process, instantly freeing gigabytes of RAM back to the operating system without waiting for the garbage collector to slowly unwind the massive AST dictionaries.