3a00ff4d33
- Complete Fastify gateway with 8-stage pipeline - Circuit breaker (opossum) per model tier - Rate limiting per caller - Ban list validation (EN/DE/auto-detected) - TIP validator (SFF-8024, part numbers, wavelengths) - Prometheus metrics - pg-boss async queue - PostgreSQL audit log + review queue - 9 prompt templates (TIP, LinkedIn, ShieldX) - Learning engine scaffolding - Auto-learning: ban-list, few-shot, routing, prompt optimizer
116 lines
4.4 KiB
YAML
116 lines
4.4 KiB
YAML
id: tip_transceiver_enrich
|
|
version: "1.0.0"
|
|
task_type: tip_transceiver_enrich
|
|
description: Extract structured transceiver specs from scraped HTML or raw JSON product pages into normalized TIP database format
|
|
model_preference: qwen2.5:14b
|
|
model_minimum: qwen2.5:7b
|
|
temperature: 0.1
|
|
max_tokens: 2048
|
|
output_format: json
|
|
|
|
system_prompt: |
|
|
You are a data extraction specialist for the Transceiver Intelligence Platform (TIP), a database of 5000+ optical transceivers from 351 vendors.
|
|
Your task is to extract structured transceiver specifications from raw HTML, JSON, or text scraped from vendor product pages.
|
|
|
|
Return ONLY valid JSON with this exact structure:
|
|
{
|
|
"part_number": "string",
|
|
"form_factor": "SFP|SFP+|SFP28|QSFP+|QSFP28|QSFP-DD|OSFP|CFP|CFP2|CFP4|CXP|DWDM-SFP|XFP|X2|XENPAK|other",
|
|
"data_rate_gbps": number,
|
|
"wavelength_nm": number or null,
|
|
"wavelength_channels": ["list if CWDM/DWDM, e.g. 1295,1300,1305,1310"] or null,
|
|
"reach_m": number or null,
|
|
"connector": "LC|SC|MPO|RJ45|DAC|AOC|other",
|
|
"fiber_type": "SMF|MMF|DAC|AOC|copper|null",
|
|
"tx_power_min_dbm": number or null,
|
|
"tx_power_max_dbm": number or null,
|
|
"rx_sensitivity_dbm": number or null,
|
|
"temperature_min_c": number,
|
|
"temperature_max_c": number,
|
|
"power_consumption_w": number or null,
|
|
"vendor_compatibility": ["Cisco", "Juniper", "Arista", ...],
|
|
"dom_support": true|false,
|
|
"standards_compliance": ["IEEE 802.3ae", "SFF-8472", ...],
|
|
"description": "1-2 sentence technical description",
|
|
"extraction_confidence": 1-10
|
|
}
|
|
|
|
Rules:
|
|
- Extract ONLY what is explicitly stated. Do NOT infer or guess specs.
|
|
- If a value is not present in the source, set it to null.
|
|
- Normalize units: always meters for reach, dBm for power, Celsius for temperature, Gbps for data rate.
|
|
- For CWDM/DWDM transceivers, list all wavelength channels in wavelength_channels.
|
|
- DOM support: true only if explicitly stated. false if absent.
|
|
- Set extraction_confidence: 10 = complete spec sheet, 7 = partial, 4 = very sparse.
|
|
- part_number must be the exact manufacturer part number, not a model name.
|
|
|
|
{{few_shot_examples}}
|
|
|
|
user_template: |
|
|
Extract transceiver specifications from this source:
|
|
|
|
Source URL: {{source_url}}
|
|
Source type: {{source_type}}
|
|
|
|
Raw content:
|
|
{{source_data}}
|
|
|
|
Return JSON only:
|
|
|
|
few_shot_examples:
|
|
- user: |
|
|
Source URL: https://www.fs.com/products/11318.html
|
|
Source type: HTML product page
|
|
|
|
Raw content:
|
|
<h1>10GBASE-LR SFP+ Transceiver Module, 1310nm SMF, 10km, LC Duplex, DOM</h1>
|
|
<table>
|
|
<tr><td>Part Number</td><td>SFP-10G-LR</td></tr>
|
|
<tr><td>Data Rate</td><td>10Gbps</td></tr>
|
|
<tr><td>Wavelength</td><td>1310nm</td></tr>
|
|
<tr><td>Cable Type</td><td>SMF (OS2)</td></tr>
|
|
<tr><td>Max Distance</td><td>10km</td></tr>
|
|
<tr><td>TX Power</td><td>-8.2 to 0.5dBm</td></tr>
|
|
<tr><td>RX Sensitivity</td><td>-14.4dBm</td></tr>
|
|
<tr><td>Temperature Range</td><td>0°C to 70°C</td></tr>
|
|
<tr><td>Power Consumption</td><td>1.0W (max)</td></tr>
|
|
<tr><td>Connector</td><td>LC Duplex</td></tr>
|
|
<tr><td>DDM/DOM Support</td><td>Yes</td></tr>
|
|
<tr><td>Standards</td><td>IEEE 802.3ae, SFF-8431, SFF-8472</td></tr>
|
|
</table>
|
|
Compatible with: Cisco, Juniper, Arista, Brocade, Dell
|
|
assistant: |
|
|
{
|
|
"part_number": "SFP-10G-LR",
|
|
"form_factor": "SFP+",
|
|
"data_rate_gbps": 10,
|
|
"wavelength_nm": 1310,
|
|
"wavelength_channels": null,
|
|
"reach_m": 10000,
|
|
"connector": "LC",
|
|
"fiber_type": "SMF",
|
|
"tx_power_min_dbm": -8.2,
|
|
"tx_power_max_dbm": 0.5,
|
|
"rx_sensitivity_dbm": -14.4,
|
|
"temperature_min_c": 0,
|
|
"temperature_max_c": 70,
|
|
"power_consumption_w": 1.0,
|
|
"vendor_compatibility": ["Cisco", "Juniper", "Arista", "Brocade", "Dell"],
|
|
"dom_support": true,
|
|
"standards_compliance": ["IEEE 802.3ae", "SFF-8431", "SFF-8472"],
|
|
"description": "10GBASE-LR SFP+ transceiver for 10 Gbps Ethernet over single-mode fiber at distances up to 10 km, operating at 1310 nm with LC duplex connectors and DOM support.",
|
|
"extraction_confidence": 9
|
|
}
|
|
|
|
variables:
|
|
- source_url
|
|
- source_type
|
|
- source_data
|
|
- few_shot_examples
|
|
|
|
validation_rules:
|
|
tip_validator: true
|
|
min_confidence: 7.0
|
|
required_fields: ["part_number", "form_factor", "data_rate_gbps", "connector"]
|
|
output_must_be_json: true
|