notkshitij/SpecForge
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

feat(retrieval): add grade matching and same-family part disambiguation.

Browse Source 
Boost scores when query grade matches standard title grade, penalize mismatches. Add part disambiguation to correctly route queries to specific standard parts (e.g., IS 12269 (Part 1) vs (Part 2)). Regenerate retrieval results with improved ranking.
This commit is contained in:
K
2026-05-04 00:20:19 +05:30
parent 28bb4ca1de
commit 3fbf91c706
3 changed files with 448 additions and 405 deletions
Download Patch File Download Diff File Expand all files Collapse all files
data/processed/retrieval_results.json
+202 -202
View File
@@ -3,50 +3,50 @@
"id": "PUB-01",
"query": "We are a small enterprise manufacturing 33 Grade Ordinary Portland Cement. Which BIS standard covers the chemical and physical requirements for our product?",
"retrieved_standards": [
"IS 8112: 1989",
"IS 269: 1989",
"IS 8043: 1991",
"IS 12269: 1987",
"IS 8042: 1989"
"IS 8112: 1989",
"IS 8042: 1989",
"IS 1489 (Part 1): 1991"
],
"details": [
{
"standard_id": "IS 8112: 1989",
"title": "43 Grade Ordinary Portland Cement",
"category": "Cement and Concrete",
"score": 1.5662,
"matched_section": "Scope"
},
{
"standard_id": "IS 269: 1989",
"title": "Ordinary Portland Cement, 33 Grade",
"category": "Cement and Concrete",
"score": 1.3095,
"score": 1.5952,
"matched_section": "Scope"
},
{
"standard_id": "IS 8043: 1991",
"title": "Hydrophobic Portland Cement",
"category": "Cement and Concrete",
"score": 1.215,
"matched_section": "Chemical Requirements"
},
{
"standard_id": "IS 12269: 1987",
"title": "53 Grade Ordinary Portland Cement",
"category": "Cement and Concrete",
"score": 1.0738,
"matched_section": "Delivery"
"score": 1.0936,
"matched_section": "Chemical Requirements"
},
{
"standard_id": "IS 8112: 1989",
"title": "43 Grade Ordinary Portland Cement",
"category": "Cement and Concrete",
"score": 1.0367,
"matched_section": "Chemical Requirements"
},
{
"standard_id": "IS 8042: 1989",
"title": "White Portland Cement",
"category": "Cement and Concrete",
"score": 0.9815,
"score": 0.8677,
"matched_section": "Degree Of Whiteness"
},
{
"standard_id": "IS 1489 (Part 1): 1991",
"title": "Portland Pozzolana Cement",
"category": "Cement and Concrete",
"score": 0.7017,
"matched_section": "Chemical Requirements"
}
],
"latency_seconds": 0.024,
"latency_seconds": 0.0255,
"expected_standards": [
"IS 269: 1989"
]
@@ -59,46 +59,46 @@
"IS 2686: 1977",
"IS 9142: 1979",
"IS 3068: 1986",
"IS 6579: 1981"
"IS 1195: 2002"
],
"details": [
{
"standard_id": "IS 383: 1970",
"title": "Coarse And Fine Aggregates From Natural Sources For Concrete",
"category": "Cement and Concrete",
"score": 1.319,
"score": 1.6,
"matched_section": "Scope"
},
{
"standard_id": "IS 2686: 1977",
"title": "Cinder As Fine Aggregates For Use In Lime Concrete",
"category": "Cement and Concrete",
"score": 1.1269,
"score": 0.9145,
"matched_section": "Scope"
},
{
"standard_id": "IS 9142: 1979",
"title": "Artificial Lightweight Aggregates For Concrete Masonry Units",
"category": "Cement and Concrete",
"score": 1.0584,
"score": 0.8367,
"matched_section": "Scope"
},
{
"standard_id": "IS 3068: 1986",
"title": "Broken Brick (Burnt Clay) Coarse Aggregate For Use In Lime Concrete",
"category": "Building Limes",
"score": 0.9841,
"score": 0.8343,
"matched_section": "General"
},
{
"standard_id": "IS 6579: 1981",
"title": "Coarse Aggregate For Water Bound Macadam",
"category": "Stones",
"score": 0.4609,
"matched_section": "Scope"
"standard_id": "IS 1195: 2002",
"title": "Bitumen Mastic For Flooring",
"category": "Floor, Wall, Roof Coverings and Finishes",
"score": 0.5439,
"matched_section": "Grades"
}
],
"latency_seconds": 0.0168,
"latency_seconds": 0.021,
"expected_standards": [
"IS 383: 1970"
]
@@ -108,10 +108,10 @@
"query": "What is the official specification for manufacturing precast concrete pipes, both with and without reinforcement, for water mains?",
"retrieved_standards": [
"IS 458: 2003",
"IS 12709: 1994",
"IS 14143: 1994",
"IS 784: 2001",
"IS 13990: 1994",
"IS 14201: 1994"
"IS 9893: 1981",
"IS 12709: 1994"
],
"details": [
{
@@ -122,35 +122,35 @@
"matched_section": "Scope"
},
{
"standard_id": "IS 12709: 1994",
"title": "Glass Fibre Reinforced Plastics (Grp) Pipes, Joints And Fittings For Use For Potable Water Supply",
"category": "Sanitary Appliances and Water Fittings",
"score": 0.5792,
"matched_section": "Workmanship"
},
{
"standard_id": "IS 14143: 1994",
"title": "Prefabricated Brick Panel And Partially Precast Concrete Joist For Flooring And Roofing",
"standard_id": "IS 784: 2001",
"title": "Prestressed Concrete Pipes",
"category": "Cement and Concrete",
"score": 0.5219,
"matched_section": "Tests"
"score": 0.6057,
"matched_section": "Scope"
},
{
"standard_id": "IS 13990: 1994",
"title": "Precast Reinforced Concrete Planks And Joists For Roofing And Flooring",
"category": "Cement and Concrete",
"score": 0.5132,
"score": 0.5667,
"matched_section": "Tests"
},
{
"standard_id": "IS 14201: 1994",
"title": "Precast Reinforced Concrete Channel Units For Construction Of Floors And Roofs",
"standard_id": "IS 9893: 1981",
"title": "Precast Concrete Blocks For Lintels And Sills",
"category": "Cement and Concrete",
"score": 0.4941,
"matched_section": "Tests"
"score": 0.5306,
"matched_section": "General"
},
{
"standard_id": "IS 12709: 1994",
"title": "Glass Fibre Reinforced Plastics (Grp) Pipes, Joints And Fittings For Use For Potable Water Supply",
"category": "Sanitary Appliances and Water Fittings",
"score": 0.5012,
"matched_section": "Workmanship"
}
],
"latency_seconds": 0.0165,
"latency_seconds": 0.018,
"expected_standards": [
"IS 458: 2003"
]
@@ -159,50 +159,50 @@
"id": "PUB-04",
"query": "Our company is shifting to manufacturing hollow and solid lightweight concrete masonry blocks. What standard outlines the dimensions and physical requirements?",
"retrieved_standards": [
"IS 2185 (Part 1): 1979",
"IS 2185 (Part 2): 1983",
"IS 2185 (Part 1): 1979",
"IS 12440: 1988",
"IS 9142: 1979",
"IS 2849: 1983"
"IS 3115: 1992",
"IS 9142: 1979"
],
"details": [
{
"standard_id": "IS 2185 (Part 1): 1979",
"title": "Concrete Masonry Units",
"category": "Cement and Concrete",
"score": 1.5,
"matched_section": "Scope"
},
{
"standard_id": "IS 2185 (Part 2): 1983",
"title": "Concrete Masonry Units",
"category": "Cement and Concrete",
"score": 1.0607,
"score": 2.4013,
"matched_section": "Scope"
},
{
"standard_id": "IS 2185 (Part 1): 1979",
"title": "Concrete Masonry Units",
"category": "Cement and Concrete",
"score": 2.3424,
"matched_section": "Scope"
},
{
"standard_id": "IS 12440: 1988",
"title": "Precast Concrete Stone Masonry Blocks",
"category": "Cement and Concrete",
"score": 0.7082,
"score": 0.9567,
"matched_section": "Classification"
},
{
"standard_id": "IS 3115: 1992",
"title": "Lime Based Blocks",
"category": "Building Limes",
"score": 0.6688,
"matched_section": "Scope"
},
{
"standard_id": "IS 9142: 1979",
"title": "Artificial Lightweight Aggregates For Concrete Masonry Units",
"category": "Cement and Concrete",
"score": 0.5147,
"matched_section": "Scope"
},
{
"standard_id": "IS 2849: 1983",
"title": "Non-Load Bearing Gypsum Partition Blocks (Solid And Hollow Types)",
"category": "Gypsum Building Materials",
"score": 0.4927,
"score": 0.5137,
"matched_section": "Scope"
}
],
"latency_seconds": 0.0161,
"latency_seconds": 0.0172,
"expected_standards": [
"IS 2185 (Part 2): 1983"
]
@@ -211,50 +211,50 @@
"id": "PUB-05",
"query": "Looking for the standard detailing corrugated and semi-corrugated asbestos cement sheets used for roofing and cladding.",
"retrieved_standards": [
"IS 13008: 1990",
"IS 13000: 1990",
"IS 459: 1992",
"IS 14241: 1995",
"IS 1626 (Part 2): 1994"
"IS 13008: 1990",
"IS 1626: 1984",
"IS 13000: 1990",
"IS 2096: 1992"
],
"details": [
{
"standard_id": "IS 13008: 1990",
"title": "Shallow Corrugated Asbestos Cement Sheets",
"standard_id": "IS 459: 1992",
"title": "Corrugated And Semi-Corrugated Asbestos",
"category": "Cement and Concrete",
"score": 1.4082,
"matched_section": "Finish"
"score": 1.4802,
"matched_section": "Scope"
},
{
"standard_id": "IS 13008: 1990",
"title": "Shallow Corrugated Asbestos",
"category": "Cement and Concrete",
"score": 1.3432,
"matched_section": "Scope"
},
{
"standard_id": "IS 1626: 1984",
"title": "Asbestos Cement Building Pipes And Pipe Fittings, Gutters And Guter Fittings And Roof Fittings",
"category": "Cement and Concrete",
"score": 0.8707,
"matched_section": "General"
},
{
"standard_id": "IS 13000: 1990",
"title": "Silica Asbestos - Cement Flat Sheets",
"category": "Cement and Concrete",
"score": 1.3157,
"score": 0.6224,
"matched_section": "Tests"
},
{
"standard_id": "IS 459: 1992",
"title": "Corrugated And Semi-Corrugated Asbestos Cement Sheets",
"standard_id": "IS 2096: 1992",
"title": "Asbestos Cement Flat Sheets",
"category": "Cement and Concrete",
"score": 1.2303,
"matched_section": "Scope"
},
{
"standard_id": "IS 14241: 1995",
"title": "Precast Reinforced Concrete L–Panel For Roofing",
"category": "Cement and Concrete",
"score": 1.0496,
"matched_section": "Scope"
},
{
"standard_id": "IS 1626 (Part 2): 1994",
"title": "Asbestos Cement Building Pipes And Pipe Fittings, Gutters And Gutter Fittings And Roof Fittings",
"category": "Cement and Concrete",
"score": 1.0191,
"score": 0.6211,
"matched_section": "Scope"
}
],
"latency_seconds": 0.0154,
"latency_seconds": 0.0169,
"expected_standards": [
"IS 459: 1992"
]
@@ -264,8 +264,8 @@
"query": "What is the Indian Standard covering the manufacture, chemical, and physical requirements for Portland slag cement?",
"retrieved_standards": [
"IS 455: 1989",
"IS 12269: 1987",
"IS 8112: 1989",
"IS 8043: 1991",
"IS 8042: 1989",
"IS 1489 (Part 1): 1991"
],
@@ -274,39 +274,39 @@
"standard_id": "IS 455: 1989",
"title": "Portland Slag Cement",
"category": "Cement and Concrete",
"score": 1.55,
"score": 1.5119,
"matched_section": "Delivery"
},
{
"standard_id": "IS 12269: 1987",
"title": "53 Grade Ordinary Portland Cement",
"category": "Cement and Concrete",
"score": 0.8,
"matched_section": "Chemical Requirements"
},
{
"standard_id": "IS 8112: 1989",
"title": "43 Grade Ordinary Portland Cement",
"category": "Cement and Concrete",
"score": 0.9617,
"matched_section": "Scope"
},
{
"standard_id": "IS 8043: 1991",
"title": "Hydrophobic Portland Cement",
"category": "Cement and Concrete",
"score": 0.9057,
"matched_section": "Chemical Requirements"
"score": 0.7961,
"matched_section": "Delivery"
},
{
"standard_id": "IS 8042: 1989",
"title": "White Portland Cement",
"category": "Cement and Concrete",
"score": 0.8542,
"score": 0.7747,
"matched_section": "Degree Of Whiteness"
},
{
"standard_id": "IS 1489 (Part 1): 1991",
"title": "Portland Pozzolana Cement",
"category": "Cement and Concrete",
"score": 0.7186,
"score": 0.713,
"matched_section": "Scope"
}
],
"latency_seconds": 0.0152,
"latency_seconds": 0.0185,
"expected_standards": [
"IS 455: 1989"
]
@@ -316,49 +316,49 @@
"query": "We are setting up a plant to produce Portland pozzolana cement that is calcined clay based. What is the applicable standard?",
"retrieved_standards": [
"IS 1489 (Part 2): 1991",
"IS 1489 (Part 1): 1991",
"IS 8112: 1989",
"IS 455: 1989",
"IS 269: 1989",
"IS 8041: 1990"
"IS 8041: 1990",
"IS 8112: 1989"
],
"details": [
{
"standard_id": "IS 1489 (Part 2): 1991",
"title": "Portland Pozzolana Cement",
"category": "Cement and Concrete",
"score": 1.4985,
"score": 3.891,
"matched_section": "Chemical Requirements"
},
{
"standard_id": "IS 1489 (Part 1): 1991",
"title": "Portland Pozzolana Cement",
"standard_id": "IS 455: 1989",
"title": "Portland Slag Cement",
"category": "Cement and Concrete",
"score": 1.4086,
"score": 0.7159,
"matched_section": "Delivery"
},
{
"standard_id": "IS 269: 1989",
"title": "Ordinary Portland Cement, 33 Grade",
"category": "Cement and Concrete",
"score": 0.4989,
"matched_section": "Scope"
},
{
"standard_id": "IS 8041: 1990",
"title": "Rapid Hardening Portland Cement",
"category": "Cement and Concrete",
"score": 0.4903,
"matched_section": "Chemical Requirements"
},
{
"standard_id": "IS 8112: 1989",
"title": "43 Grade Ordinary Portland Cement",
"category": "Cement and Concrete",
"score": 0.5554,
"matched_section": "Scope"
},
{
"standard_id": "IS 269: 1989",
"title": "Ordinary Portland Cement, 33 Grade",
"category": "Cement and Concrete",
"score": 0.5504,
"matched_section": "Delivery"
},
{
"standard_id": "IS 8041: 1990",
"title": "Rapid Hardening Portland Cement",
"category": "Cement and Concrete",
"score": 0.5152,
"matched_section": "Physical Requirements"
"score": 0.3611,
"matched_section": "Chemical Requirements"
}
],
"latency_seconds": 0.0174,
"latency_seconds": 0.0181,
"expected_standards": [
"IS 1489 (Part 2): 1991"
]
@@ -367,50 +367,50 @@
"id": "PUB-08",
"query": "Which standard applies to masonry cement used for general purposes where mortars for masonry are required, but not intended for structural concrete?",
"retrieved_standards": [
"IS 1489 (Part 2): 1991",
"IS 3466: 1988",
"IS 2185 (Part 1): 1979",
"IS 2185 (Part 3): 1984",
"IS 2185 (Part 1): 1979",
"IS 2116: 1980",
"IS 2185 (Part 2): 1983"
],
"details": [
{
"standard_id": "IS 1489 (Part 2): 1991",
"title": "Portland Pozzolana Cement",
"category": "Cement and Concrete",
"score": 1.1,
"matched_section": "Scope"
},
{
"standard_id": "IS 3466: 1988",
"title": "Masonry Cement",
"category": "Cement and Concrete",
"score": 0.722,
"matched_section": "Scope"
},
{
"standard_id": "IS 2185 (Part 1): 1979",
"title": "Concrete Masonry Units",
"category": "Cement and Concrete",
"score": 0.6658,
"score": 1.45,
"matched_section": "Scope"
},
{
"standard_id": "IS 2185 (Part 3): 1984",
"title": "Concrete Masonry Units",
"category": "Cement and Concrete",
"score": 0.6417,
"score": 0.6326,
"matched_section": "Scope"
},
{
"standard_id": "IS 2185 (Part 1): 1979",
"title": "Concrete Masonry Units",
"category": "Cement and Concrete",
"score": 0.6167,
"matched_section": "Scope"
},
{
"standard_id": "IS 2116: 1980",
"title": "Sand For Masonry Mortars",
"category": "Cement and Concrete",
"score": 0.5807,
"matched_section": "General"
},
{
"standard_id": "IS 2185 (Part 2): 1983",
"title": "Concrete Masonry Units",
"category": "Cement and Concrete",
"score": 0.6017,
"score": 0.5712,
"matched_section": "Classification"
}
],
"latency_seconds": 0.0167,
"latency_seconds": 0.0209,
"expected_standards": [
"IS 3466: 1988"
]
@@ -420,49 +420,49 @@
"query": "Looking for the standard that details the composition, manufacture, and testing of supersulphated cement, particularly for marine works or aggressive water conditions.",
"retrieved_standards": [
"IS 6909: 1990",
"IS 6452: 1989",
"IS 12269: 1987",
"IS 13114: 1991",
"IS 778: 1984"
"IS 6163: 1978",
"IS 4984: 1995",
"IS 6598: 1972"
],
"details": [
{
"standard_id": "IS 6909: 1990",
"title": "Supersulphated Cement",
"category": "Cement and Concrete",
"score": 1.05,
"matched_section": "Chemical Requirements"
},
{
"standard_id": "IS 6452: 1989",
"title": "High Alumina Cement For Structural Use",
"category": "Cement and Concrete",
"score": 1.0179,
"matched_section": "Applications"
},
{
"standard_id": "IS 12269: 1987",
"title": "53 Grade Ordinary Portland Cement",
"category": "Cement and Concrete",
"score": 0.5367,
"matched_section": "Scope"
"score": 0.9,
"matched_section": "General"
},
{
"standard_id": "IS 13114: 1991",
"title": "Forged Brass Gate, Globe And Check Valves For Water Works Purposes",
"category": "Building Limes",
"score": 0.4346,
"score": 0.6,
"matched_section": "Dimensions And Tolerances"
},
{
"standard_id": "IS 778: 1984",
"title": "Copper Alloy Gate, Globe And Check Valves For Water Works Purposes",
"category": "Building Limes",
"score": 0.4274,
"matched_section": "Design And Manufacture"
"standard_id": "IS 6163: 1978",
"title": "Centrifugally Cast (Spun) Iron Low Pressure Pipes For Water, Gas And Sewage",
"category": "Sanitary Appliances and Water Fittings",
"score": 0.4635,
"matched_section": "Scope"
},
{
"standard_id": "IS 4984: 1995",
"title": "High Density Polyethylene Pipes For Water Supply",
"category": "Sanitary Appliances and Water Fittings",
"score": 0.4291,
"matched_section": "Composition"
},
{
"standard_id": "IS 6598: 1972",
"title": "Cellular Concrete For Thermal Insulation",
"category": "Thermal Insulation Materials",
"score": 0.4215,
"matched_section": "Scope"
}
],
"latency_seconds": 0.0178,
"latency_seconds": 0.0191,
"expected_standards": [
"IS 6909: 1990"
]
@@ -473,48 +473,48 @@
"retrieved_standards": [
"IS 8042: 1989",
"IS 8112: 1989",
"IS 8043: 1991",
"IS 455: 1989",
"IS 1489 (Part 1): 1991"
"IS 8041: 1990",
"IS 12269: 1987",
"IS 455: 1989"
],
"details": [
{
"standard_id": "IS 8042: 1989",
"title": "White Portland Cement",
"category": "Cement and Concrete",
"score": 1.3082,
"score": 1.3724,
"matched_section": "Degree Of Whiteness"
},
{
"standard_id": "IS 8112: 1989",
"title": "43 Grade Ordinary Portland Cement",
"category": "Cement and Concrete",
"score": 1.1737,
"score": 1.022,
"matched_section": "Scope"
},
{
"standard_id": "IS 8043: 1991",
"title": "Hydrophobic Portland Cement",
"standard_id": "IS 8041: 1990",
"title": "Rapid Hardening Portland Cement",
"category": "Cement and Concrete",
"score": 0.9227,
"score": 0.8907,
"matched_section": "Chemical Requirements"
},
{
"standard_id": "IS 12269: 1987",
"title": "53 Grade Ordinary Portland Cement",
"category": "Cement and Concrete",
"score": 0.8447,
"matched_section": "Scope"
},
{
"standard_id": "IS 455: 1989",
"title": "Portland Slag Cement",
"category": "Cement and Concrete",
"score": 0.7604,
"matched_section": "Delivery"
},
{
"standard_id": "IS 1489 (Part 1): 1991",
"title": "Portland Pozzolana Cement",
"category": "Cement and Concrete",
"score": 0.7296,
"score": 0.7568,
"matched_section": "Delivery"
}
],
"latency_seconds": 0.0156,
"latency_seconds": 0.0159,
"expected_standards": [
"IS 8042: 1989"
]
data/results.json
+200 -200
View File
@@ -3,50 +3,50 @@
"id": "PUB-01",
"query": "We are a small enterprise manufacturing 33 Grade Ordinary Portland Cement. Which BIS standard covers the chemical and physical requirements for our product?",
"retrieved_standards": [
"IS 8112: 1989",
"IS 269: 1989",
"IS 8043: 1991",
"IS 12269: 1987",
"IS 8042: 1989"
"IS 8112: 1989",
"IS 8042: 1989",
"IS 1489 (Part 1): 1991"
],
"details": [
{
"standard_id": "IS 8112: 1989",
"title": "43 Grade Ordinary Portland Cement",
"category": "Cement and Concrete",
"score": 1.5662,
"matched_section": "Scope"
},
{
"standard_id": "IS 269: 1989",
"title": "Ordinary Portland Cement, 33 Grade",
"category": "Cement and Concrete",
"score": 1.3095,
"score": 1.5952,
"matched_section": "Scope"
},
{
"standard_id": "IS 8043: 1991",
"title": "Hydrophobic Portland Cement",
"category": "Cement and Concrete",
"score": 1.215,
"matched_section": "Chemical Requirements"
},
{
"standard_id": "IS 12269: 1987",
"title": "53 Grade Ordinary Portland Cement",
"category": "Cement and Concrete",
"score": 1.0738,
"matched_section": "Delivery"
"score": 1.0936,
"matched_section": "Chemical Requirements"
},
{
"standard_id": "IS 8112: 1989",
"title": "43 Grade Ordinary Portland Cement",
"category": "Cement and Concrete",
"score": 1.0367,
"matched_section": "Chemical Requirements"
},
{
"standard_id": "IS 8042: 1989",
"title": "White Portland Cement",
"category": "Cement and Concrete",
"score": 0.9815,
"score": 0.8677,
"matched_section": "Degree Of Whiteness"
},
{
"standard_id": "IS 1489 (Part 1): 1991",
"title": "Portland Pozzolana Cement",
"category": "Cement and Concrete",
"score": 0.7017,
"matched_section": "Chemical Requirements"
}
],
"latency_seconds": 0.0221,
"latency_seconds": 0.0237,
"expected_standards": [
"IS 269: 1989"
]
@@ -59,46 +59,46 @@
"IS 2686: 1977",
"IS 9142: 1979",
"IS 3068: 1986",
"IS 6579: 1981"
"IS 1195: 2002"
],
"details": [
{
"standard_id": "IS 383: 1970",
"title": "Coarse And Fine Aggregates From Natural Sources For Concrete",
"category": "Cement and Concrete",
"score": 1.319,
"score": 1.6,
"matched_section": "Scope"
},
{
"standard_id": "IS 2686: 1977",
"title": "Cinder As Fine Aggregates For Use In Lime Concrete",
"category": "Cement and Concrete",
"score": 1.1269,
"score": 0.9145,
"matched_section": "Scope"
},
{
"standard_id": "IS 9142: 1979",
"title": "Artificial Lightweight Aggregates For Concrete Masonry Units",
"category": "Cement and Concrete",
"score": 1.0584,
"score": 0.8367,
"matched_section": "Scope"
},
{
"standard_id": "IS 3068: 1986",
"title": "Broken Brick (Burnt Clay) Coarse Aggregate For Use In Lime Concrete",
"category": "Building Limes",
"score": 0.9841,
"score": 0.8343,
"matched_section": "General"
},
{
"standard_id": "IS 6579: 1981",
"title": "Coarse Aggregate For Water Bound Macadam",
"category": "Stones",
"score": 0.4609,
"matched_section": "Scope"
"standard_id": "IS 1195: 2002",
"title": "Bitumen Mastic For Flooring",
"category": "Floor, Wall, Roof Coverings and Finishes",
"score": 0.5439,
"matched_section": "Grades"
}
],
"latency_seconds": 0.0202,
"latency_seconds": 0.0201,
"expected_standards": [
"IS 383: 1970"
]
@@ -108,10 +108,10 @@
"query": "What is the official specification for manufacturing precast concrete pipes, both with and without reinforcement, for water mains?",
"retrieved_standards": [
"IS 458: 2003",
"IS 12709: 1994",
"IS 14143: 1994",
"IS 784: 2001",
"IS 13990: 1994",
"IS 14201: 1994"
"IS 9893: 1981",
"IS 12709: 1994"
],
"details": [
{
@@ -122,32 +122,32 @@
"matched_section": "Scope"
},
{
"standard_id": "IS 12709: 1994",
"title": "Glass Fibre Reinforced Plastics (Grp) Pipes, Joints And Fittings For Use For Potable Water Supply",
"category": "Sanitary Appliances and Water Fittings",
"score": 0.5792,
"matched_section": "Workmanship"
},
{
"standard_id": "IS 14143: 1994",
"title": "Prefabricated Brick Panel And Partially Precast Concrete Joist For Flooring And Roofing",
"standard_id": "IS 784: 2001",
"title": "Prestressed Concrete Pipes",
"category": "Cement and Concrete",
"score": 0.5219,
"matched_section": "Tests"
"score": 0.6057,
"matched_section": "Scope"
},
{
"standard_id": "IS 13990: 1994",
"title": "Precast Reinforced Concrete Planks And Joists For Roofing And Flooring",
"category": "Cement and Concrete",
"score": 0.5132,
"score": 0.5667,
"matched_section": "Tests"
},
{
"standard_id": "IS 14201: 1994",
"title": "Precast Reinforced Concrete Channel Units For Construction Of Floors And Roofs",
"standard_id": "IS 9893: 1981",
"title": "Precast Concrete Blocks For Lintels And Sills",
"category": "Cement and Concrete",
"score": 0.4941,
"matched_section": "Tests"
"score": 0.5306,
"matched_section": "General"
},
{
"standard_id": "IS 12709: 1994",
"title": "Glass Fibre Reinforced Plastics (Grp) Pipes, Joints And Fittings For Use For Potable Water Supply",
"category": "Sanitary Appliances and Water Fittings",
"score": 0.5012,
"matched_section": "Workmanship"
}
],
"latency_seconds": 0.0176,
@@ -159,50 +159,50 @@
"id": "PUB-04",
"query": "Our company is shifting to manufacturing hollow and solid lightweight concrete masonry blocks. What standard outlines the dimensions and physical requirements?",
"retrieved_standards": [
"IS 2185 (Part 1): 1979",
"IS 2185 (Part 2): 1983",
"IS 2185 (Part 1): 1979",
"IS 12440: 1988",
"IS 9142: 1979",
"IS 2849: 1983"
"IS 3115: 1992",
"IS 9142: 1979"
],
"details": [
{
"standard_id": "IS 2185 (Part 1): 1979",
"title": "Concrete Masonry Units",
"category": "Cement and Concrete",
"score": 1.5,
"matched_section": "Scope"
},
{
"standard_id": "IS 2185 (Part 2): 1983",
"title": "Concrete Masonry Units",
"category": "Cement and Concrete",
"score": 1.0607,
"score": 2.4013,
"matched_section": "Scope"
},
{
"standard_id": "IS 2185 (Part 1): 1979",
"title": "Concrete Masonry Units",
"category": "Cement and Concrete",
"score": 2.3424,
"matched_section": "Scope"
},
{
"standard_id": "IS 12440: 1988",
"title": "Precast Concrete Stone Masonry Blocks",
"category": "Cement and Concrete",
"score": 0.7082,
"score": 0.9567,
"matched_section": "Classification"
},
{
"standard_id": "IS 3115: 1992",
"title": "Lime Based Blocks",
"category": "Building Limes",
"score": 0.6688,
"matched_section": "Scope"
},
{
"standard_id": "IS 9142: 1979",
"title": "Artificial Lightweight Aggregates For Concrete Masonry Units",
"category": "Cement and Concrete",
"score": 0.5147,
"matched_section": "Scope"
},
{
"standard_id": "IS 2849: 1983",
"title": "Non-Load Bearing Gypsum Partition Blocks (Solid And Hollow Types)",
"category": "Gypsum Building Materials",
"score": 0.4927,
"score": 0.5137,
"matched_section": "Scope"
}
],
"latency_seconds": 0.0194,
"latency_seconds": 0.019,
"expected_standards": [
"IS 2185 (Part 2): 1983"
]
@@ -211,50 +211,50 @@
"id": "PUB-05",
"query": "Looking for the standard detailing corrugated and semi-corrugated asbestos cement sheets used for roofing and cladding.",
"retrieved_standards": [
"IS 13008: 1990",
"IS 13000: 1990",
"IS 459: 1992",
"IS 14241: 1995",
"IS 1626 (Part 2): 1994"
"IS 13008: 1990",
"IS 1626: 1984",
"IS 13000: 1990",
"IS 2096: 1992"
],
"details": [
{
"standard_id": "IS 13008: 1990",
"title": "Shallow Corrugated Asbestos Cement Sheets",
"standard_id": "IS 459: 1992",
"title": "Corrugated And Semi-Corrugated Asbestos",
"category": "Cement and Concrete",
"score": 1.4082,
"matched_section": "Finish"
"score": 1.4802,
"matched_section": "Scope"
},
{
"standard_id": "IS 13008: 1990",
"title": "Shallow Corrugated Asbestos",
"category": "Cement and Concrete",
"score": 1.3432,
"matched_section": "Scope"
},
{
"standard_id": "IS 1626: 1984",
"title": "Asbestos Cement Building Pipes And Pipe Fittings, Gutters And Guter Fittings And Roof Fittings",
"category": "Cement and Concrete",
"score": 0.8707,
"matched_section": "General"
},
{
"standard_id": "IS 13000: 1990",
"title": "Silica Asbestos - Cement Flat Sheets",
"category": "Cement and Concrete",
"score": 1.3157,
"score": 0.6224,
"matched_section": "Tests"
},
{
"standard_id": "IS 459: 1992",
"title": "Corrugated And Semi-Corrugated Asbestos Cement Sheets",
"standard_id": "IS 2096: 1992",
"title": "Asbestos Cement Flat Sheets",
"category": "Cement and Concrete",
"score": 1.2303,
"matched_section": "Scope"
},
{
"standard_id": "IS 14241: 1995",
"title": "Precast Reinforced Concrete L–Panel For Roofing",
"category": "Cement and Concrete",
"score": 1.0496,
"matched_section": "Scope"
},
{
"standard_id": "IS 1626 (Part 2): 1994",
"title": "Asbestos Cement Building Pipes And Pipe Fittings, Gutters And Gutter Fittings And Roof Fittings",
"category": "Cement and Concrete",
"score": 1.0191,
"score": 0.6211,
"matched_section": "Scope"
}
],
"latency_seconds": 0.0184,
"latency_seconds": 0.0181,
"expected_standards": [
"IS 459: 1992"
]
@@ -264,8 +264,8 @@
"query": "What is the Indian Standard covering the manufacture, chemical, and physical requirements for Portland slag cement?",
"retrieved_standards": [
"IS 455: 1989",
"IS 12269: 1987",
"IS 8112: 1989",
"IS 8043: 1991",
"IS 8042: 1989",
"IS 1489 (Part 1): 1991"
],
@@ -274,35 +274,35 @@
"standard_id": "IS 455: 1989",
"title": "Portland Slag Cement",
"category": "Cement and Concrete",
"score": 1.55,
"score": 1.5119,
"matched_section": "Delivery"
},
{
"standard_id": "IS 12269: 1987",
"title": "53 Grade Ordinary Portland Cement",
"category": "Cement and Concrete",
"score": 0.8,
"matched_section": "Chemical Requirements"
},
{
"standard_id": "IS 8112: 1989",
"title": "43 Grade Ordinary Portland Cement",
"category": "Cement and Concrete",
"score": 0.9617,
"matched_section": "Scope"
},
{
"standard_id": "IS 8043: 1991",
"title": "Hydrophobic Portland Cement",
"category": "Cement and Concrete",
"score": 0.9057,
"matched_section": "Chemical Requirements"
"score": 0.7961,
"matched_section": "Delivery"
},
{
"standard_id": "IS 8042: 1989",
"title": "White Portland Cement",
"category": "Cement and Concrete",
"score": 0.8542,
"score": 0.7747,
"matched_section": "Degree Of Whiteness"
},
{
"standard_id": "IS 1489 (Part 1): 1991",
"title": "Portland Pozzolana Cement",
"category": "Cement and Concrete",
"score": 0.7186,
"score": 0.713,
"matched_section": "Scope"
}
],
@@ -316,49 +316,49 @@
"query": "We are setting up a plant to produce Portland pozzolana cement that is calcined clay based. What is the applicable standard?",
"retrieved_standards": [
"IS 1489 (Part 2): 1991",
"IS 1489 (Part 1): 1991",
"IS 8112: 1989",
"IS 455: 1989",
"IS 269: 1989",
"IS 8041: 1990"
"IS 8041: 1990",
"IS 8112: 1989"
],
"details": [
{
"standard_id": "IS 1489 (Part 2): 1991",
"title": "Portland Pozzolana Cement",
"category": "Cement and Concrete",
"score": 1.4985,
"score": 3.891,
"matched_section": "Chemical Requirements"
},
{
"standard_id": "IS 1489 (Part 1): 1991",
"title": "Portland Pozzolana Cement",
"standard_id": "IS 455: 1989",
"title": "Portland Slag Cement",
"category": "Cement and Concrete",
"score": 1.4086,
"score": 0.7159,
"matched_section": "Delivery"
},
{
"standard_id": "IS 269: 1989",
"title": "Ordinary Portland Cement, 33 Grade",
"category": "Cement and Concrete",
"score": 0.4989,
"matched_section": "Scope"
},
{
"standard_id": "IS 8041: 1990",
"title": "Rapid Hardening Portland Cement",
"category": "Cement and Concrete",
"score": 0.4903,
"matched_section": "Chemical Requirements"
},
{
"standard_id": "IS 8112: 1989",
"title": "43 Grade Ordinary Portland Cement",
"category": "Cement and Concrete",
"score": 0.5554,
"matched_section": "Scope"
},
{
"standard_id": "IS 269: 1989",
"title": "Ordinary Portland Cement, 33 Grade",
"category": "Cement and Concrete",
"score": 0.5504,
"matched_section": "Delivery"
},
{
"standard_id": "IS 8041: 1990",
"title": "Rapid Hardening Portland Cement",
"category": "Cement and Concrete",
"score": 0.5152,
"matched_section": "Physical Requirements"
"score": 0.3611,
"matched_section": "Chemical Requirements"
}
],
"latency_seconds": 0.0188,
"latency_seconds": 0.02,
"expected_standards": [
"IS 1489 (Part 2): 1991"
]
@@ -367,50 +367,50 @@
"id": "PUB-08",
"query": "Which standard applies to masonry cement used for general purposes where mortars for masonry are required, but not intended for structural concrete?",
"retrieved_standards": [
"IS 1489 (Part 2): 1991",
"IS 3466: 1988",
"IS 2185 (Part 1): 1979",
"IS 2185 (Part 3): 1984",
"IS 2185 (Part 1): 1979",
"IS 2116: 1980",
"IS 2185 (Part 2): 1983"
],
"details": [
{
"standard_id": "IS 1489 (Part 2): 1991",
"title": "Portland Pozzolana Cement",
"category": "Cement and Concrete",
"score": 1.1,
"matched_section": "Scope"
},
{
"standard_id": "IS 3466: 1988",
"title": "Masonry Cement",
"category": "Cement and Concrete",
"score": 0.722,
"matched_section": "Scope"
},
{
"standard_id": "IS 2185 (Part 1): 1979",
"title": "Concrete Masonry Units",
"category": "Cement and Concrete",
"score": 0.6658,
"score": 1.45,
"matched_section": "Scope"
},
{
"standard_id": "IS 2185 (Part 3): 1984",
"title": "Concrete Masonry Units",
"category": "Cement and Concrete",
"score": 0.6417,
"score": 0.6326,
"matched_section": "Scope"
},
{
"standard_id": "IS 2185 (Part 1): 1979",
"title": "Concrete Masonry Units",
"category": "Cement and Concrete",
"score": 0.6167,
"matched_section": "Scope"
},
{
"standard_id": "IS 2116: 1980",
"title": "Sand For Masonry Mortars",
"category": "Cement and Concrete",
"score": 0.5807,
"matched_section": "General"
},
{
"standard_id": "IS 2185 (Part 2): 1983",
"title": "Concrete Masonry Units",
"category": "Cement and Concrete",
"score": 0.6017,
"score": 0.5712,
"matched_section": "Classification"
}
],
"latency_seconds": 0.0197,
"latency_seconds": 0.0193,
"expected_standards": [
"IS 3466: 1988"
]
@@ -420,49 +420,49 @@
"query": "Looking for the standard that details the composition, manufacture, and testing of supersulphated cement, particularly for marine works or aggressive water conditions.",
"retrieved_standards": [
"IS 6909: 1990",
"IS 6452: 1989",
"IS 12269: 1987",
"IS 13114: 1991",
"IS 778: 1984"
"IS 6163: 1978",
"IS 4984: 1995",
"IS 6598: 1972"
],
"details": [
{
"standard_id": "IS 6909: 1990",
"title": "Supersulphated Cement",
"category": "Cement and Concrete",
"score": 1.05,
"matched_section": "Chemical Requirements"
},
{
"standard_id": "IS 6452: 1989",
"title": "High Alumina Cement For Structural Use",
"category": "Cement and Concrete",
"score": 1.0179,
"matched_section": "Applications"
},
{
"standard_id": "IS 12269: 1987",
"title": "53 Grade Ordinary Portland Cement",
"category": "Cement and Concrete",
"score": 0.5367,
"matched_section": "Scope"
"score": 0.9,
"matched_section": "General"
},
{
"standard_id": "IS 13114: 1991",
"title": "Forged Brass Gate, Globe And Check Valves For Water Works Purposes",
"category": "Building Limes",
"score": 0.4346,
"score": 0.6,
"matched_section": "Dimensions And Tolerances"
},
{
"standard_id": "IS 778: 1984",
"title": "Copper Alloy Gate, Globe And Check Valves For Water Works Purposes",
"category": "Building Limes",
"score": 0.4274,
"matched_section": "Design And Manufacture"
"standard_id": "IS 6163: 1978",
"title": "Centrifugally Cast (Spun) Iron Low Pressure Pipes For Water, Gas And Sewage",
"category": "Sanitary Appliances and Water Fittings",
"score": 0.4635,
"matched_section": "Scope"
},
{
"standard_id": "IS 4984: 1995",
"title": "High Density Polyethylene Pipes For Water Supply",
"category": "Sanitary Appliances and Water Fittings",
"score": 0.4291,
"matched_section": "Composition"
},
{
"standard_id": "IS 6598: 1972",
"title": "Cellular Concrete For Thermal Insulation",
"category": "Thermal Insulation Materials",
"score": 0.4215,
"matched_section": "Scope"
}
],
"latency_seconds": 0.0183,
"latency_seconds": 0.0229,
"expected_standards": [
"IS 6909: 1990"
]
@@ -473,48 +473,48 @@
"retrieved_standards": [
"IS 8042: 1989",
"IS 8112: 1989",
"IS 8043: 1991",
"IS 455: 1989",
"IS 1489 (Part 1): 1991"
"IS 8041: 1990",
"IS 12269: 1987",
"IS 455: 1989"
],
"details": [
{
"standard_id": "IS 8042: 1989",
"title": "White Portland Cement",
"category": "Cement and Concrete",
"score": 1.3082,
"score": 1.3724,
"matched_section": "Degree Of Whiteness"
},
{
"standard_id": "IS 8112: 1989",
"title": "43 Grade Ordinary Portland Cement",
"category": "Cement and Concrete",
"score": 1.1737,
"score": 1.022,
"matched_section": "Scope"
},
{
"standard_id": "IS 8043: 1991",
"title": "Hydrophobic Portland Cement",
"standard_id": "IS 8041: 1990",
"title": "Rapid Hardening Portland Cement",
"category": "Cement and Concrete",
"score": 0.9227,
"score": 0.8907,
"matched_section": "Chemical Requirements"
},
{
"standard_id": "IS 12269: 1987",
"title": "53 Grade Ordinary Portland Cement",
"category": "Cement and Concrete",
"score": 0.8447,
"matched_section": "Scope"
},
{
"standard_id": "IS 455: 1989",
"title": "Portland Slag Cement",
"category": "Cement and Concrete",
"score": 0.7604,
"matched_section": "Delivery"
},
{
"standard_id": "IS 1489 (Part 1): 1991",
"title": "Portland Pozzolana Cement",
"category": "Cement and Concrete",
"score": 0.7296,
"score": 0.7568,
"matched_section": "Delivery"
}
],
"latency_seconds": 0.0197,
"latency_seconds": 0.0182,
"expected_standards": [
"IS 8042: 1989"
]
inference.py
+46 -3
View File
@@ -54,6 +54,10 @@ _TOP_N_FINAL = 5
_SHORT_CHUNK_THRESHOLD = 40 # body words below this get a penalty
_SHORT_CHUNK_PENALTY = 0.15
_GRADE_RE = re.compile(r"\b(33|43|53)\b")
_PART_BASE_RE = re.compile(r"IS\s+(\d+)\s*\(", re.IGNORECASE)
_STOP_WORDS = {"and", "or", "for", "the", "of", "in", "a", "an", "to"}
# ---------------------------------------------------------------------------
# Text helpers
@@ -205,6 +209,9 @@ class Retriever:
query_lower = query.lower()
query_words = set(query_tokens)
grade_in_query = set(_GRADE_RE.findall(query))
sig_query = query_words - _STOP_WORDS
for cid, base in list(chunk_scores.items()):
chunk = self.idx.chunks[cid]
sid_norm = _norm_std_id(chunk["standard_id"])
@@ -227,9 +234,7 @@ class Retriever:
# Strong boost: majority of title words present in query — likely
# the most on-point standard even if its chunk body is polluted.
stop = {"and", "or", "for", "the", "of", "in", "a", "an", "to"}
sig_title = full_title_tokens - stop
sig_query = query_words - stop
sig_title = full_title_tokens - _STOP_WORDS
if sig_title and len(sig_title & sig_query) / len(sig_title) >= 0.6:
bonus += 0.25
@@ -245,6 +250,13 @@ class Retriever:
if body_wc < _SHORT_CHUNK_THRESHOLD:
bonus -= _SHORT_CHUNK_PENALTY
if grade_in_query:
grade_in_title = set(_GRADE_RE.findall(full_title))
if grade_in_title and grade_in_title == grade_in_query:
bonus += 0.35
elif grade_in_title and grade_in_title.isdisjoint(grade_in_query):
bonus -= 0.40
chunk_scores[cid] = base + bonus
# --- Group by standard_id, keep best chunk score ---
@@ -257,6 +269,37 @@ class Retriever:
std_best[sid] = score
std_chunk_repr[sid] = chunk
# --- Same-family Part disambiguation ---
base_to_sids: dict[str, list[str]] = {}
for sid in std_best:
if "(" not in sid:
continue
m = _PART_BASE_RE.match(sid)
if m:
base_to_sids.setdefault(m.group(1), []).append(sid)
for sids in base_to_sids.values():
if len(sids) < 2:
continue
# Only act when all siblings share the same title
titles = {_norm_std_id(std_chunk_repr[s].get("full_title") or
std_chunk_repr[s].get("title", ""))
for s in sids}
if len(titles) > 1:
continue
sib_kws = {s: self.idx.std_keywords.get(_norm_std_id(s), set()) for s in sids}
disc_weight: dict[str, float] = {s: 0.0 for s in sids}
for token in query_words:
owners = [s for s in sids if token in sib_kws[s]]
if len(owners) == 1:
# IDF floors at 0 for very common terms; +1 keeps weight positive
idf_weight = max(self.idx.bm25.idf.get(token, 0.0), 0.0) + 1.0
disc_weight[owners[0]] += idf_weight
max_w = max(disc_weight.values())
if max_w == 0.0:
continue
for s, w in disc_weight.items():
std_best[s] += w * 0.22 - (max_w - w) * 0.14
# --- Sort and take top N ---
ranked = sorted(std_best.items(), key=lambda x: x[1], reverse=True)[:top_n]