constraint-table-post-processing-reasoning

name

constraint / table / post-processing reasoning

description

Use this skill when the user wants table-heavy questions, multi-constraint mathematical logic, SQL/programming tracing, or questions requiring post-processing formats. Trigger it for requests like 'make it use a table and text', 'trace the math logic of this textbook', 'ask for the one that fits several conditions', 'cross-reference the table with the article', or 'force an algorithmic calculation'. This skill targets complex operation execution after foundational text retrieval.

Skill: constraint / table / post-processing reasoning

1. Capability Definition & Real Case

• Professional Definition: The ability to accurately synthesize an answer by subjecting retrieved evidence to explicit secondary post-processing operations. This encompasses structured matrix calculations, mapping text-based entities into relational database executions (Text-to-SQL logic), mathematical/algorithmic formula tracing, and intersectional constraint satisfaction. The mechanism separates finding raw factual information from executing strict operational structures upon it.
• Dimension Hierarchy: Multi-Source Evidence Composition->Cross-Document Synthesis->constraint / table / post-processing reasoning

Real Case

[Case 1]

• Initial Environment: A bounded sports-information corpus explicitly containing structured tables or infobox-like grids. Facts reside strictly in row-column junctions.
• Real Question: How many runs did the West Indies vice-captain score in the 1983 World Cup?
• Real Trajectory: 1. Extract the text 'West Indies vice-captain' and resolve it to 'Viv Richards' using contextual lookup. 2. Transition to analyzing the structured table data for the 1983 tournament. 3. Intersect the 'Viv Richards' row constraint with the 'Runs' column constraint. 4. Output the numeric cell data.
• Real Answer: A numeric answer extracted from table-like evidence.
• Why this demonstrates the capability: The challenge demands spatial alignment rather than paragraph scanning. The agent actively binds an identified team role trait with separate tournament data across disparate axes to compute a targeted intersection inside grid parameters.

---

[Case 2]

• Initial Environment: A mixed-modality corpus encompassing a 'Billboard 200' historical text document and a separate, rigorous structured database outlining 2013 chart statistics.
• Real Question: Which record labels released the album that hit number one on the Billboard leaderboard in 2013?
• Real Trajectory: 1. Analyze structured database schema for year 2013 targeting the 'Rank 1' entity. 2. Extract table output result 'ArtPop'. 3. Transition logic to semantic search within unstructured prose documents targeting 'ArtPop publisher'. 4. Combine the SQL-derived entity with the text-derived label array.
• Real Answer: Streamline and Interscope Records
• Why this demonstrates the capability: Achieving the correct answer demands heterogeneous source tracking. The agent operates symbol-based filters on structured tables to unlock textual constraints applied against unstructured data paragraphs, forging a multi-modal reasoning pipeline.

---

[Case 3]

• Initial Environment: A technical textbook repository concentrating exclusively on Computer Vision architecture, holding separate passages indicating mathematical scaling formulas and distinct scenario parameters.
• Real Question: Given an input volume of 227x227x3, a Convolutional layer with 96 filters of size 11x11, a stride of 4, and no padding, followed by a Max Pooling layer of size 3x3 with a stride of 2, what are the final output dimensions?
• Real Trajectory: 1. Search technical manual texts identifying abstract pooling formulas (Output = (W-F+2P)/S+1). 2. Map query parameters to layer 1 math constraints, yielding 55. 3. Input first-hop resolution directly into step 2 math equations yielding 27. 4. Merge mathematical layer iterations into sequential formatting rules.
• Real Answer: 27x27x96
• Why this demonstrates the capability: Success originates specifically from identifying overarching architectural formulas stripped from abstract settings, then rigorously feeding textually-constrained scenario metrics through iterative sequential math proofs independent of standard string retrieval mechanisms.

Pipeline Execution Instructions

To synthesize data for this capability, you must strictly follow a 3-phase pipeline. Do not hallucinate steps. Read the corresponding reference file for each phase sequentially:

1. Phase 1: Environment Exploration Read the exploration guidelines to discover raw knowledge seeds: references/EXPLORATION.md

2. Phase 2: Trajectory Selection Once Phase 1 is complete, read the selection criteria to evaluate the trajectory: references/SELECTION.md

3. Phase 3: Data Synthesis Once a trajectory passes Phase 2, read the synthesis instructions to generate the final data: references/SYNTHESIS.md