Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/apursley2012/rescue-retriever/llms.txt

Use this file to discover all available pages before exploring further.

Rescue Retriever supports three distinct search-and-rescue program profiles — Water Rescue, Mountain or Wilderness Rescue, and Disaster Rescue or Individual Tracking. Each profile defines its own qualifying breeds, required sex, and eligible age range. Rather than producing a binary pass/fail result, the scoring system assigns a numeric score to every animal in the dataset for each program, allowing shelter coordinators and rescue organization representatives to rank all animals by fit and prioritize follow-up accordingly.

​
Scoring Formula

Every animal receives a score between 22 and 100 for each rescue program profile. The score is the sum of three independent components — breed, sex, and age — capped at 100.
CriterionMatchNo Match
Breed50 pts12 pts
Sex25 pts5 pts
Age25 pts5 pts
  • Minimum possible score: 22 (no criterion met: 12 + 5 + 5)
  • Maximum possible score: 100 (all criteria met: 50 + 25 + 25)
score = min(100,
  (breedMatch ? 50 : 12) +
  (sexMatch   ? 25 :  5) +
  (ageMatch   ? 25 :  5)
)
Strict qualification requires all three criteria to be satisfied simultaneously. An animal with a score of 100 that fails any single criterion is not considered a qualified candidate — qualifies is only true when breedMatch, sexMatch, and ageMatch are all true. The numeric score remains useful for ranking partial matches and surfacing near-miss animals for human review.

​
Programs at a Glance

Water Rescue

Targets Intact Female retrieving breeds aged 26–156 weeks for aquatic search operations. 87% training success rate.

Mountain Rescue

Targets Intact Male cold-climate working breeds aged 26–156 weeks for avalanche and backcountry operations. 82% training success rate.

Disaster Rescue

Targets Intact Male elite tracking breeds aged 20–300 weeks for urban rubble and individual scent tracking. 91% training success rate.

​
Breed Matching

Breed matching uses normalized substring comparison rather than exact string equality. Before comparison, both the animal’s breed field and each breed in the profile’s breed list are processed through the same normalization steps:
  1. Convert to lowercase
  2. Strip the words mixed and dog
  3. Collapse any resulting extra whitespace
After normalization, a match is recorded if either string contains the other as a substring. This bidirectional check means that a shelter record reading “Labrador Retriever Mix” will successfully match the profile breed “Labrador Retriever Mix” even after both sides are normalized. It also allows partial shelter classifications — for example, a record entered as “Chesapeake Bay” — to still resolve to the correct profile breed.

​
Age Handling

Age values in the Austin Animal Center dataset are stored as floating-point weeks. The application rounds each animal’s age to the nearest whole number using Math.round() before applying the profile’s minAgeWeeks and maxAgeWeeks range check. The same rounded value is used for all display purposes throughout the dashboard. An animal with an ageWeeks value of 25.6 rounds to 26 and would be within the Water Rescue age window; an animal at 25.4 rounds to 25 and would fall outside it.
Breed classification in shelter records is a visual estimate made at intake, not a result of genetic testing. Studies have shown that visual breed identification agrees with DNA panel results in as few as 25–30% of cases for mixed-breed animals. Rescue Retriever surfaces candidates for human review — final selection always requires in-person evaluation by a qualified rescue program representative.

Build docs developers (and LLMs) love

Get started for freeTalk to us