Upload folder using huggingface_hub

chatbot/chunks/wiki_chunks.json

@@ -167,6 +167,147 @@
     "page_type": "entity",
     "layer": "wiki"
   },
+  {
+    "text": "[\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\" — Inventory of CSA TIMPs for Maize Value Chain — Volume I]\n**Citation:** Musila R.N., Ligeyo D.O., Murenga M. et al. (2022). *Inventory of Climate Smart Agriculture Technologies, Innovations and Management Practices for Maize Value Chain, Volume I.* KALRO/KCSAP. October 2022.",
+    "source": "wiki/sources/maize-timps-volume1.md",
+    "title": "\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\"",
+    "section": "Inventory of CSA TIMPs for Maize Value Chain — Volume I",
+    "page_type": "source-summary",
+    "layer": "wiki"
+  },
+  {
+    "text": "[\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\" — Overview]\nThis 365-page document is the official KCSAP TIMPs inventory for the maize value chain. It provides standardized, structured profiles of **193 TIMPs** covering the full production system from variety selection through postharvest and value addition.\n\n**Totals:** 116 Technologies | 20 Innovations | 57 Management Practices  \n**Status:** 149 ready for upscaling | 37 requires validation | 5 requires further research\n\nEach TIMP follows a uniform 7-section template: (A) problem/description/justification, (B) dissemination approaches, (C) counties promoted/to be upscaled, (D) economic/gender/VMG considerations, (E) success stories, (F) readiness status, (G) contacts.\n\n---",
+    "source": "wiki/sources/maize-timps-volume1.md",
+    "title": "\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\"",
+    "section": "Overview",
+    "page_type": "source-summary",
+    "layer": "wiki"
+  },
+  {
+    "text": "[\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\" — Sub-theme Summary]\n| Sub-theme | Technologies | Innovations | Mgmt Practices |\n|---|---|---|---|\n| Improved maize varieties | 79 | 0 | 0 |\n| Maize seed system | 0 | 3 | 0 |\n| GAPs and Food Safety | 1 | 0 | 2 |\n| Agronomic management practices | 0 | 0 | 9 |\n| Soil fertility management | 0 | 1 | 3 |\n| Soil & water management | 9 | 0 | 2 |\n| Irrigation and drainage | 1 | 0 | 0 |\n| Agroforestry systems | 0 | 0 | 1 |\n| Maize Crop health | 3 | 0 | 15 |\n| Weed Management | 2 | 1 | 9 |\n| Harvesting & Postharvest | 10 | 0 | 3 |\n| Maize Value addition | 3 | 14 | 0 |\n| Mechanization | 8 | 1 | 0 |\n| Farming Business & marketing | 0 | 0 | 8 |\n| Elements of Agricultural Policies | 0 | 0 | 5 |\n\n---",
+    "source": "wiki/sources/maize-timps-volume1.md",
+    "title": "\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\"",
+    "section": "Sub-theme Summary",
+    "page_type": "source-summary",
+    "layer": "wiki"
+  },
+  {
+    "text": "[\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\" — 2.2.1 Coastal Lowlands (CL 2, 3 & 4)]\n| TIMP Name | Type | Maturity | Yield potential | Key attributes | Status |\n|---|---|---|---|---|---|\n| Coast Composite Maize (CCM) | OPV | 110–120 days | 16–21 bags/acre | Heat tolerant, leaf rust resistant, rainfed+irrigated | Ready |\n| WSQ104 | QPM OPV | 90–105 days | 15 bags/acre | Drought tolerant, high lysine+tryptophan, 0–1200 m | Ready |\n| KH500Q | QPM 3-way hybrid | 90–120 days | 36 bags/acre | Drought tolerant, MSV resistant, GLS resistant | Ready |\n| PH4 (Pwani Hybrid 4) | Hybrid | 120–150 days | 24 bags/acre | Heat tolerant, partial MSV resistance, rainfed+irrigated | Ready |\n| PH1 (Pwani Hybrid 1) | Hybrid | 90–120 days | 18 bags/acre | Drought tolerant, excellent husk cover | Ready |\n| WE2111 | 3-way hybrid | 4.5–5 months | 4.7–8.7 t/ha | Drought tolerant, NLB/GLS/MSV resistant, white dent | Ready |\n| MTPEH0701 | 3-way hybrid | 120–150 days | 26 bags/acre | Large grain borer & weevil resistant | Ready |\n| MTPEH0702 | 3-way hybrid | 120–150 days | 26 bags/acre | Large grain borer & weevil resistant | Ready |\n| MTPEH0703 | 3-way hybrid | 4–5 months | 26 bags/acre | Spotted stem borer resistant | Ready |\n| MTPEH200804 (KH125-02-MDR) | 3-way hybrid | 120 days | 26 bags/acre | MSV resistant, GLS resistant | Ready |\n| MTPEH200805 (KH125-03-SG) | Stay-green 3-way hybrid | 120 days | 5–6.4 t/ha | Drought tolerant, stem borer resistant, GLS resistant | Ready |",
+    "source": "wiki/sources/maize-timps-volume1.md",
+    "title": "\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\"",
+    "section": "2.2.1 Coastal Lowlands (CL 2, 3 & 4)",
+    "page_type": "source-summary",
+    "layer": "wiki"
+  },
+  {
+    "text": "[\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\" — 2.2.2 Medium Altitude – Dry (semi-arid, ~600–1200 m)]\n| TIMP Name | Type | Maturity | Yield potential | Key attributes | Status |\n|---|---|---|---|---|---|\n| KCB (Katumani Composite B) | OPV | 85–95 days | 16–21 bags/acre | Drought escaping, stem borer resistant | Ready |\n| KDH6 SBR | 3-way hybrid | 90–120 days | 5 t/ha | Drought tolerant, low-N tolerant, stem borer resistant | Ready |\n| KH414-03 SBR | 3-way hybrid | 90–120 days | 4 t/ha | Stem borer resistant | Ready |\n| KDH414-11 (Ukamez 6) | 3-way hybrid | 90–100 days | 4.6–7.5 t/ha | Drought tolerant, GLS/NLB/MSV resistant | Ready |\n| KDH414-12 (Ukamez 7) | Stay-green 3-way hybrid | 90–100 days | 4.3–7.8 t/ha | Drought tolerant, GLS/NLB/MSV resistant, good for livestock feed | Ready |\n| WE2109 | 3-way hybrid | 4.5–5 months | 4.8–9.2 t/ha | Drought tolerant, NLB/GLS/MSV resistant, white dent | Ready |",
+    "source": "wiki/sources/maize-timps-volume1.md",
+    "title": "\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\"",
+    "section": "2.2.2 Medium Altitude – Dry (semi-arid, ~600–1200 m)",
+    "page_type": "source-summary",
+    "layer": "wiki"
+  },
+  {
+    "text": "[\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\" — 2.2.3 Medium Altitude – Moist (~1000–1800 m)]\n| TIMP Name | Type | Maturity | Yield potential | Key attributes | Status |\n|---|---|---|---|---|---|\n| WE2101 | 3-way hybrid | 4.5 months | 6.9 t/ha | Drought tolerant, NLB/MSV/GLS resistant, good husk cover | Ready |\n| WE2104 | 3-way hybrid | 4.5 months | 7.16 t/ha | Drought tolerant, NLB/MSV/GLS resistant | Ready |\n| WE2107 | 3-way hybrid | 4.5 months | 7 t/ha | NLB/MSV/GLS resistant, good husk cover | Ready |\n| WE2108 | 3-way hybrid | 4.5 months | 6.9 t/ha | NLB/MSV/GLS resistant, good husk cover | Ready |\n| WE5206 | 3-way hybrid | 105–130 days | 9 t/ha | Drought tolerant, NLB/MSV/GLS resistant | Ready |\n| WE5230 | 3-way hybrid | 105–130 days | 8 t/ha | Drought tolerant, NLB/MSV/GLS resistant | Ready |\n| EMB225 (KBEST) | 3-way hybrid | 90–120 days | 4.6 t/ha | NLB/MSV/leaf rust resistant, 1000–1800 m | Ready |\n| EMB226 (EMBU POA) | Stay-green 3-way hybrid | 90–120 days | 6.5 t/ha | NLB resistant, 1200–1800 m | Ready |\n| KH500-40E | 3-way hybrid | 120–130 days | 7 t/ha | Drought tolerant, low-N tolerant | Ready |\n| KH500-39E | 3-way hybrid | 120 days | 8–10 t/ha | Drought tolerant, GLS/NLB/MSV/foliar resistant | Ready |\n| KH500-56A (KM1101) | 3-way hybrid | 5–6 months | 6.5 t/ha | MSV resistant, GLS resistant, good husk cover | Ready |\n| KH500-Q | 3-way hybrid | 120 days | 8–10 t/ha | QPM, drought tolerant, GLS/NLB/MSV resistant | Ready |",
+    "source": "wiki/sources/maize-timps-volume1.md",
+    "title": "\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\"",
+    "section": "2.2.3 Medium Altitude – Moist (~1000–1800 m)",
+    "page_type": "source-summary",
+    "layer": "wiki"
+  },
+  {
+    "text": "[\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\" — 2.2.4 Medium Altitude – Transitional]\n| TIMP Name | Type | Maturity | Yield potential | Key attributes | Status |\n|---|---|---|---|---|---|\n| WE2106 | 3-way hybrid | 4.5–5 months | 4.7–9.1 t/ha | Drought tolerant, NLB/GLS/MSV resistant, good husk cover | Ready |\n| WE3104 | 3-way hybrid | 4 months | 5.6 t/ha | NLB/MSV/GLS resistant, good husk cover | Ready |\n| WE3106 | 3-way hybrid | 4 months | 3.28 t/ha | NLB/MSV/GLS resistant, good husk cover | Ready |\n| WE5107 | 3-way hybrid | 105–130 days | 3.7–7.2 t/ha | Drought tolerant, NLB/GLS/MSV resistant | Ready |\n| WE5113 | 3-way hybrid | 105–130 days | 3.7–7.3 t/ha | Drought tolerant, NLB/GLS/MSV resistant | Ready |\n| WE5138 | 3-way hybrid | 105–130 days | 4–6.5 t/ha | Drought tolerant, NLB/GLS/MSV resistant | Ready |\n| WE5205 | 3-way hybrid | 105–130 days | 7.4–9.6 t/ha | Drought tolerant, NLB/GLS/MSV resistant | Ready |\n| WE5210 | 3-way hybrid | 105–130 days | 7.6–10.0 t/ha | Drought tolerant, NLB/GLS/MSV resistant, good husk cover | Ready |\n| WE5213 | 3-way hybrid | 105–120 days | 7.5–9.6 t/ha | Drought tolerant, NLB/GLS/MSV resistant, good husk cover | Ready |\n| WE5218 | 3-way hybrid | 105–125 days | 7.3–9.1 t/ha | Drought tolerant, NLB/GLS",
+    "source": "wiki/sources/maize-timps-volume1.md",
+    "page": 0,
+    "layer": "wiki",
+    "title": "\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\"",
+    "section": "2.2.4 Medium Altitude – Transitional",
+    "page_type": "source-summary"
+  },
+  {
+    "text": "/GLS/MSV resistant, good husk cover | Ready |\n| WE5218 | 3-way hybrid | 105–125 days | 7.3–9.1 t/ha | Drought tolerant, NLB/GLS/MSV resistant | Ready |\n| WE5227 | 3-way hybrid | 105–130 days | 7.5–9.5 t/ha | Drought tolerant, NLB/GLS/MSV resistant | Ready |\n| KH500-13E | 3-way hybrid | 120 days | 4.5–5.3 t/ha | Drought tolerant, ear rot resistant, NLB/GLS/MSV resistant | Ready |\n| KH500-51A (MU07-018) | 3-way hybrid | 150–170 days | 6.0–6.5 t/ha | NLB/GLS/MSV resistant, flint-like | Ready |\n| KH500-52A (MU08-005) | 3-way hybrid | 4–5 months | 4.39 t/ha | Drought tolerant, NLB/GLS/MSV resistant | Ready |\n| KH500-53A (MU10-010) | 3-way hybrid | 150–170 days | 4.9–5.0 t/ha | Drought tolerant, NLB/GLS/MSV resistant | Ready |\n| KH500-54A (CKH10778) | 3-way hybrid | 5–6 months | 8.5 t/ha | MSV resistant, GLS resistant, good husk cover | Ready |\n| KH500-55A (MU10-233) | 3-way hybrid | 120–160 days | 6.5 t/ha | Drought tolerant, MSV resistant, GLS resistant | Ready |",
+    "source": "wiki/sources/maize-timps-volume1.md",
+    "page": 0,
+    "layer": "wiki",
+    "title": "\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\"",
+    "section": "2.2.4 Medium Altitude – Transitional",
+    "page_type": "source-summary"
+  },
+  {
+    "text": "[\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\" — 2.2.5 Highlands (1800–2500 m asl, 1000–2000 mm rainfall)]\n| TIMP Name | Type | Maturity | Yield potential | Key attributes | Status |\n|---|---|---|---|---|---|\n| KH600-11D | Varietal cross hybrid | 140–160 days | 32–42 bags/acre | GLS resistant, strong stalks, good husk cover (1999) | Ready |\n| KH600-23A | Top cross hybrid | 145–175 days | 43–68 bags/acre | GLS/rust resistant, strong stalks, good husk cover (2008) | Ready |\n| HAC (High Altitude Composite) | OPV | 140–160 days | 20–34 bags/acre | Early maturity, frost tolerant, GLS resistant, 2200–3000 m (2006) | Ready |\n| KH600-14E | Top cross hybrid | 150–165 days | 38–48 bags/acre | Drought tolerant, GLS/rust/TLB resistant (2004) | Ready |\n| KH600-15A | Top cross hybrid | 145–180 days | 33–47 bags/acre | GLS/rust/blight resistant, strong stalks (2001) | Ready |\n| KH600-16A | Top cross hybrid | 140–180 days | 35–48 bags/acre | MSV/GLS/rust/blight resistant (2001) | Ready |\n| KH600-17A | Varietal cross hybrid | 140–160 days | 37–51 bags/acre | GLS/blight resistant, strong stalks (2002) | Ready |\n| KH600-18A | Varietal hybrid | 155–170 days | 36–50 bags/acre | GLS/blight resistant, strong stalks (2004) | Ready |\n| KH600-19A | Double cross hybrid | 160–175 days | 38–53 bags/acre | GLS/blight resistant, strong stalks (2005) | Ready |\n| KH600-20A | Top cross hybrid | 160–180 days | 38–",
+    "source": "wiki/sources/maize-timps-volume1.md",
+    "page": 0,
+    "layer": "wiki",
+    "title": "\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\"",
+    "section": "2.2.5 Highlands (1800–2500 m asl, 1000–2000 mm rainfall)",
+    "page_type": "source-summary"
+  },
+  {
+    "text": " 160–175 days | 38–53 bags/acre | GLS/blight resistant, strong stalks (2005) | Ready |\n| KH600-20A | Top cross hybrid | 160–180 days | 38–55 bags/acre | GLS/rust/blight resistant, strong stalks, good husk cover (2005) | Ready |\n| KH600-21A | Top cross hybrid | 140–180 days | 38–62 bags/acre | GLS/rust/blight resistant, strong stalks, good husk cover (2006) | Ready |\n| KH600-22A | Top cross hybrid | 140–180 days | 40–66 bags/acre | GLS resistant, good husk cover, strong stalks (2006) | Ready |\n| KH600-24A | Top cross hybrid | 140–160 days | 47–69 bags/acre | GLS/rust/blight resistant, strong roots, good husk cover (2001) | Ready |\n| KH600-25A | Top cross hybrid | 140–165 days | 49–70 bags/acre | GLS/rust/blight resistant, good husk cover (2015) | Ready |\n| KH600-26A | Top cross hybrid | 142–170 days | 51–70 bags/acre | GLS/rust/blight resistant, good husk cover, strong stalks (2015) | Ready |\n| KH600-27A | Top cross hybrid | 145–175 days | 53–71 bags/acre | GLS/rust/blight resistant, good husk cover, strong stalks (2015) | Ready |\n| H614D | Top cross hybrid | 145–175 days | 30–35 bags/acre | Drought tolerant, GLS/rust/blight resistant (1986); Kenya Seed Co. | Ready |\n| H626 | Double cross hybrid | 160–180 days | 38–42 bags/acre | Rust/blight resistant, strong stalks, good husk cover (1989) | Ready |\n| H625 | Double cross hybrid | 155–180 days | 37–40 bags/acre | Rust/blight resistant, strong stalks, good husk cover (1981) | Ready |\n| H6213 | Double cross hybrid | 160–180 days",
+    "source": "wiki/sources/maize-timps-volume1.md",
+    "page": 0,
+    "layer": "wiki",
+    "title": "\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\"",
+    "section": "2.2.5 Highlands (1800–2500 m asl, 1000–2000 mm rainfall)",
+    "page_type": "source-summary"
+  },
+  {
+    "text": " hybrid | 155–180 days | 37–40 bags/acre | Rust/blight resistant, strong stalks, good husk cover (1981) | Ready |\n| H6213 | Double cross hybrid | 160–180 days | 52–56 bags/acre | GLS/rust/blight resistant, good husk cover (2002) | Ready |\n| H6218 | Double cross hybrid | 145–175 days | 56–71 bags/acre | GLS/rust/blight resistant, good husk cover (2004) | Ready |\n| H6210 | Double cross hybrid | 160–185 days | 50–53 bags/acre | GLS/rust/blight resistant, good husk cover (2004) | Ready |\n| H629 | Double cross hybrid | 160–175 days | 48–52 bags/acre | GLS/rust/blight resistant, good husk cover (2000) | Ready |\n| H628 | Double hybrid | 155–175 days | 46–50 bags/acre | GLS/rust/blight resistant, good husk cover (1999) | Ready |\n| H624 | Double cross hybrid | 135–150 days | 30–32 bags/acre | GLS/rust/blight resistant, early maturity (2004); 1600–2300 m | Ready |\n\n**Note:** H614D, H626, H625, H6213, H6218, H6210, H629, H628, H624 are licensed to **Kenya Seed Company** for basic and commercial seed production. KH600 series breeder seed is at KALRO-Kitale.",
+    "source": "wiki/sources/maize-timps-volume1.md",
+    "page": 0,
+    "layer": "wiki",
+    "title": "\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\"",
+    "section": "2.2.5 Highlands (1800–2500 m asl, 1000–2000 mm rainfall)",
+    "page_type": "source-summary"
+  },
+  {
+    "text": "[\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\" — 2.2.6 Special Kit – MLN Tolerant Varieties]\n| TIMP Name | Type | Key attributes | Status |\n|---|---|---|---|\n| WE5135 | 3-way hybrid | MLN tolerant (MCMV+SCMV) | Ready |\n| WE5139 | 3-way hybrid | MLN tolerant | Ready |\n| WE5140 | 3-way hybrid | MLN tolerant | Ready |\n| KATEH16-02 | 3-way hybrid | MLN tolerant | Ready |\n| KATEH16-03 | 3-way hybrid | MLN tolerant | Ready |\n\n---",
+    "source": "wiki/sources/maize-timps-volume1.md",
+    "title": "\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\"",
+    "section": "2.2.6 Special Kit – MLN Tolerant Varieties",
+    "page_type": "source-summary",
+    "layer": "wiki"
+  },
+  {
+    "text": "[\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\" — 2.3 Maize Seed System]\n| TIMP Name | Category | Status |\n|---|---|---|\n| Improved Farmer-Saved-Seed-System | Innovation | Requires validation |\n| Quality Declared Seed System (QDS) | Innovation | Requires validation |\n| Maize Formal Seed System | Innovation | Ready for upscaling |\n\n---",
+    "source": "wiki/sources/maize-timps-volume1.md",
+    "title": "\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\"",
+    "section": "2.3 Maize Seed System",
+    "page_type": "source-summary",
+    "layer": "wiki"
+  },
+  {
+    "text": "[\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\" — 2.4 GAPs and Food Safety]\n| TIMP Name | Category | Status |\n|---|---|---|\n| Food Safety Management System (HACCP) | Management practice | Ready for upscaling |\n| Good Agricultural Practices (GAPs) for maize | Management practice | Ready for upscaling |\n| Aflasafe KE01™ | Technology | Requires validation |\n\n**Aflasafe KE01™**: Pre-harvest bio-control agent that reduces aflatoxin contamination in maize by 80–99% at harvest and in storage. Applied 2–3 weeks before flowering at 10 kg/ha. Manufactured at KALRO-Katumani, distributed by Koppert Biological Systems. Cost: KES 201/kg.\n\n---",
+    "source": "wiki/sources/maize-timps-volume1.md",
+    "title": "\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\"",
+    "section": "2.4 GAPs and Food Safety",
+    "page_type": "source-summary",
+    "layer": "wiki"
+  },
+  {
+    "text": "[\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\" — 2.5 Agronomic Management Practices (all Ready for upscaling)]\n| TIMP | Key recommendation |\n|---|---|\n| Land preparation | One tractor plough + one harrow; or two animal draft ploughs |\n| Maize variety selection | Match variety to AEZ; use certified seed |\n| Planting spacing | Highland: 75×25 cm (1 plant); Medium: 75×30 cm; Dryland/coastal: 90×30 cm |\n| Intercropping | Maize–legume intercrop improves soil fertility and weed suppression |\n| Weeding | Hand weed 3 weeks after emergence; repeat 3 weeks later |\n| Basal fertilizer | 50 kg/acre DAP (neutral soils) or 100 kg/acre NPK 23:23:0 (acid soils) |\n| Top-dressing fertilizer | 50 kg/acre CAN at 8–10 leaf stage (45 cm height) |\n| Timely harvesting | Harvest at physiological maturity (black layer, leaves dried) |\n| Crop rotation | Rotate with legumes to replenish N; avoid same family consecutively |\n\n---",
+    "source": "wiki/sources/maize-timps-volume1.md",
+    "title": "\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\"",
+    "section": "2.5 Agronomic Management Practices (all Ready for upscaling)",
+    "page_type": "source-summary",
+    "layer": "wiki"
+  },
+  {
+    "text": "[\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\" — 2.6 Soil Fertility Management]\n| TIMP | Category | Status |\n|---|---|---|\n| Integrated Soil Fertility Management (ISFM) | Management practice | Requires validation |\n| Integrated Manure Management (IMM) | Management practice | Requires further research |\n| Rapid soil testing services (spectroscopy/Soil Cares) | Innovation | Requires validation |\n| Low-Cost Composting technology | Management practice | Requires validation |\n\n---",
+    "source": "wiki/sources/maize-timps-volume1.md",
+    "title": "\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\"",
+    "section": "2.6 Soil Fertility Management",
+    "page_type": "source-summary",
+    "layer": "wiki"
+  },
+  {
+    "text": "[\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\" — Key Cross-References]\n- Varieties for MLN-endemic areas → [[maize-lethal-necrosis]], section 2.2.6\n- Water harvesting in ASAL → [[water-harvesting-technologies]] (zai pits, contour bunds in sections 2.7)\n- KCSAP counties targeted → [[KCSAP]] (24 counties)\n- Seed system innovations → [[seed-systems]]\n- Lead station for highland varieties → KALRO-Kitale (FCRI, Centre Director D.O. Ligeyo)\n- Lead station for coastal varieties → KALRO-Mtwapa / KALRO-Katumani\n- Lead station for mid-altitude varieties → KALRO-Embu, KALRO-Katumani",
+    "source": "wiki/sources/maize-timps-volume1.md",
+    "title": "\"Inventory of CSA Technologies, Innovations and Management Practices for Maize Value Chain (KALRO/KCSAP, 2022)\"",
+    "section": "Key Cross-References",
+    "page_type": "source-summary",
+    "layer": "wiki"
+  },
   {
     "text": "[\"Tied Ridges, FYM, and Nitrogen on Soil Moisture and Maize Yield in Machakos (Ngie Mwende, 2019)\" — Tied Ridges, FYM, and Nitrogen on Soil Moisture and Maize Yield in Machakos]\n**Citation:** Ngie Mwende, M. (2019). *Effect of tied ridges, farmyard manure, nitrogen fertilizer, and cropping systems on soil moisture, soil properties, and maize yield in Machakos County.* PhD thesis, Kenyatta University.",
     "source": "wiki/sources/tied-ridges-soil-moisture-machakos.md",

chatbot/extracted/_extraction_summary.json

@@ -34,6 +34,11 @@
     "status": "ok",
     "pages": 10
   },
+  {
+    "file": "MAIZE_TIMPS_Volume-1_for_Upload.pdf",
+    "status": "ok",
+    "pages": 372
+  },
   {
     "file": "Maize production systems  farmers  perception and current status of maize lethal necrosis in selected counties in Kenya.pdf",
     "status": "ok",

chatbot/pipeline/chunk_timps.py

@@ -0,0 +1,212 @@
+"""
+TIMP-aware chunker for MAIZE_TIMPS_Volume-1_for_Upload.json.
+
+Splits on numbered section boundaries (2.2.1.1, 2.3.2, etc.) and sub-splits
+each TIMP at its A/B/C/D/E/F/G section labels when the block is too large.
+Metadata headers are prepended so retrieval can filter by AEZ / category.
+
+Run standalone or imported by chunk.py.
+"""
+
+import json
+import re
+import tiktoken
+from pathlib import Path
+
+EXTRACTED_DIR = Path(__file__).parent.parent / "extracted"
+TIMPS_FILE = EXTRACTED_DIR / "MAIZE_TIMPS_Volume-1_for_Upload.json"
+
+CHUNK_TOKENS = 500
+OVERLAP_TOKENS = 50
+enc = tiktoken.get_encoding("cl100k_base")
+
+# ── AEZ / theme mapping ────────────────────────────────────────────────────────
+
+AEZ_MAP = {
+    "2.2.1": "Coastal Lowlands",
+    "2.2.2": "Medium Altitude – Dry",
+    "2.2.3": "Medium Altitude – Moist",
+    "2.2.4": "Medium Altitude – Transitional",
+    "2.2.5": "Highlands",
+    "2.2.6": "MLN-Tolerant (Special Kit)",
+    "2.3":   "Maize Seed System",
+    "2.4":   "Good Agricultural Practices",
+    "2.5":   "Agronomic Management Practices",
+    "2.6":   "Soil Fertility Management",
+    "2.7":   "Water Harvesting Technologies",
+    "2.8":   "Pest and Disease Management",
+    "2.9":   "Post-Harvest Management",
+    "2.10":  "Market Linkages",
+}
+
+CATEGORY_MAP = {
+    "2.2": "Improved Maize Variety",
+    "2.3": "Seed System Innovation",
+    "2.4": "Good Agricultural Practice",
+    "2.5": "Agronomic Management Practice",
+    "2.6": "Soil Fertility Management",
+    "2.7": "Water Harvesting Technology",
+    "2.8": "Pest/Disease Management",
+    "2.9": "Post-Harvest Management",
+    "2.10": "Market Linkage",
+}
+
+SOURCE = "MAIZE_TIMPS_Volume-1_for_Upload.pdf"
+TOC_CUTOFF = 15000  # characters — skip text before this as table-of-contents
+
+
+# ── helpers ────────────────────────────────────────────────────────────────────
+
+def count_tokens(text: str) -> int:
+    return len(enc.encode(text))
+
+
+def sliding_window(text: str, meta_header: str, chunk_size: int, overlap: int) -> list[str]:
+    """Split text into overlapping token windows, each prefixed with meta_header."""
+    tokens = enc.encode(text)
+    results = []
+    start = 0
+    while start < len(tokens):
+        end = min(start + chunk_size, len(tokens))
+        results.append(meta_header + enc.decode(tokens[start:end]))
+        if end == len(tokens):
+            break
+        start += chunk_size - overlap
+    return results
+
+
+def get_aez(section_code: str) -> str:
+    parts = section_code.split(".")
+    for depth in (4, 3, 2):
+        key = ".".join(parts[:depth])
+        if key in AEZ_MAP:
+            return AEZ_MAP[key]
+    return "Unknown"
+
+
+def get_category(section_code: str) -> str:
+    parts = section_code.split(".")
+    for depth in (3, 2):
+        key = ".".join(parts[:depth])
+        if key in CATEGORY_MAP:
+            return CATEGORY_MAP[key]
+    return "TIMP"
+
+
+def extract_timp_name(block_text: str) -> str:
+    """Pull the value from the 'TIMP Name' row in a table block."""
+    m = re.search(r'TIMP [Nn]ame\s*\n(.+)', block_text)
+    if m:
+        return m.group(1).strip()
+    return ""
+
+
+def split_at_sections(block_text: str) -> list[tuple[str, str]]:
+    """
+    Split a TIMP block at its A:/B:/C:/D:/E:/F:/G: section boundaries.
+    Returns list of (label, text) pairs.
+    """
+    pattern = re.compile(
+        r'(?:^|\n)([A-G][:\.])\s*(?:Description|Assessment|Current situation|Economic|Case stud|Status|Contacts)[^\n]*\n',
+        re.IGNORECASE
+    )
+    hits = list(pattern.finditer(block_text))
+    if not hits:
+        return [("", block_text)]
+
+    parts = []
+    # Text before first section label
+    pre = block_text[:hits[0].start()].strip()
+    if pre:
+        parts.append(("header", pre))
+
+    for i, m in enumerate(hits):
+        label = m.group(1).rstrip(":.")
+        start = m.start()
+        end = hits[i + 1].start() if i + 1 < len(hits) else len(block_text)
+        parts.append((label, block_text[start:end].strip()))
+
+    return parts
+
+
+# ── main chunker ───────────────────────────────────────────────────────────────
+
+def chunk_timps() -> list[dict]:
+    if not TIMPS_FILE.exists():
+        return []
+
+    data = json.loads(TIMPS_FILE.read_text(encoding="utf-8"))
+    full_text = "\n".join(p["text"] for p in sorted(data["pages"], key=lambda x: x["page"]))
+
+    # Find all section boundary positions (skip TOC area).
+    # Match only known theme prefixes (2.2–2.10), avoiding table row artifacts like 2.1.1.
+    sec_pattern = re.compile(
+        r"(?:^|\n)(2\.(?:[2-9]|10)(?:\.\d+){1,2})\s+([^\n.]{3,80})\n",
+        re.MULTILINE,
+    )
+    boundaries = [
+        (m.start(), m.group(1), m.group(2).strip())
+        for m in sec_pattern.finditer(full_text)
+        if m.start() > TOC_CUTOFF
+    ]
+
+    if not boundaries:
+        return []
+
+    chunks: list[dict] = []
+
+    for idx, (pos, code, title) in enumerate(boundaries):
+        # Block text runs until the next boundary (or end of document)
+        next_pos = boundaries[idx + 1][0] if idx + 1 < len(boundaries) else len(full_text)
+        block = full_text[pos:next_pos].strip()
+
+        aez = get_aez(code)
+        category = get_category(code)
+        timp_name = extract_timp_name(block) or title
+
+        meta_header = (
+            f"[TIMP: {timp_name}]\n"
+            f"[Category: {category}]\n"
+            f"[AEZ/Theme: {aez}]\n"
+            f"[Section: {code}]\n\n"
+        )
+
+        # Sub-split at A-G section boundaries to keep chunks focused
+        section_parts = split_at_sections(block)
+
+        for label, part_text in section_parts:
+            if not part_text.strip():
+                continue
+            full_chunk = meta_header + (f"Section {label}:\n" if label else "") + part_text
+            if count_tokens(full_chunk) <= CHUNK_TOKENS * 1.5:
+                chunks.append({
+                    "text": full_chunk,
+                    "source": SOURCE,
+                    "layer": "pdf",
+                    "timp_name": timp_name,
+                    "aez": aez,
+                    "category": category,
+                    "section_code": code,
+                    "section_label": label,
+                })
+            else:
+                for window in sliding_window(part_text, meta_header, CHUNK_TOKENS, OVERLAP_TOKENS):
+                    chunks.append({
+                        "text": window,
+                        "source": SOURCE,
+                        "layer": "pdf",
+                        "timp_name": timp_name,
+                        "aez": aez,
+                        "category": category,
+                        "section_code": code,
+                        "section_label": label,
+                    })
+
+    return chunks
+
+
+if __name__ == "__main__":
+    result = chunk_timps()
+    print(f"TIMP chunks: {len(result)}")
+    if result:
+        print(f"Sample chunk:\n{result[0]['text'][:300]}")