内容
[](https://pypi.org/project/aion-indian-market-intelligence/)
# AION Indian Market Intelligence for Macro Event & Sector Impact Analysis
**An API that turns an Indian financial headline into a signed sector vector —
which sectors are exposed, in which direction, and who is affected.**
A sentiment score says `negative 0.82`. Negative *for whom?* One event moves
different sectors in opposite directions, and a single number has to pick one and
discard the rest.
## What it does
```mermaid
flowchart LR
A["Your headline feed<br/>NSE circulars · news · RSS"] --> B["analyze(headline)"]
B --> C["AION IMI<br/>sector_vector + context"]
C --> D["Your LLM · agent · dashboard"]
D --> E["Prioritised actions<br/>risk flags · alerts · research"]
```
## Quickstart in 60 seconds
```bash
pip install aion-indian-market-intelligence
export AION_API_KEY="<your_api_key>"
```
```python
from aion_indian_market_intelligence import analyze
intel = analyze("RBI hikes repo rate by 25 bps")
print(intel["sector_vector"])
# {'Information Technology': 0.18, 'Realty': -0.52,
# 'Financial Services': -0.31, 'Automobile and Auto Components': -0.26}
```
Get a key at
[dashboard.aiondashboard.site/access/register](https://dashboard.aiondashboard.site/access/register).
The free tier is 1,000 requests a month.
## What this is NOT
> - **Not investment advice.** It describes exposure; it does not tell you what to do about it.
> - **Not a trading signal.** There is no entry, exit, target or position size anywhere in the output.
> - **Not a downloadable model.** Inference runs through the hosted API. This repository contains no weights, coefficients, taxonomy files or server code.
## Before / After
The same headline, through a polarity model and through this one:
> Unseasonal rainfall and hailstorm hit apple orchards in Himachal Pradesh in April
| | Polarity-only NLP | AION IMI |
| --- | --- | --- |
| **Output** | `negative` | `event: rain_apple_damage` |
| **Which sectors** | — | Agriculture & Horticulture, Transportation, Consumer Services under pressure |
| **Flip side** | — | storage-linked and substitute supply chains may benefit |
| **Who is affected** | — | producer losses · policymaker inflation watch · investor attention on second-order effects |
| **Actionable by a system** | not really | yes — it names sectors you can join to your own data |
Market implication: agriculture-linked exposure under pressure, logistics
bottlenecks possible, substitute supply chains may benefit.
## Use Cases
Three patterns, all of which people are already building. Longer write-ups in
[the DEV.to article](https://dev.to/aion-analytics-india/aion-indian-market-intelligence-the-india-contextual-intelligence-layer-for-your-llms-and-saas-4eam).
### 1. LLM / RAG augmentation
*For AI engineers.* Store the sector vector next to the headline in your vector
store. When the LLM retrieves the document it also retrieves India-specific
impact context, so it reasons about the event rather than paraphrasing it.
```python
intel = analyze(headline)
enriched = (
f"Headline: {headline}\n"
f"Sector impact: {intel['sector_vector']}\n"
f"Stakeholders: {intel.get('stakeholder_views')}"
)
# embed `enriched`, or store the vector as metadata alongside the headline
```
### 2. SaaS sales & CRM intelligence
*For B2B SaaS founders.* Map a macro event onto your own customer list by sector.
Instead of "this is negative", your system knows which accounts sit in the
affected sectors, in which direction, and how hard.
```mermaid
flowchart TD
E["Policy · weather · commodity shock"] --> I["analyze() → sector impact + severity"]
I --> C["Your CRM / lead list, by sector"]
C --> L["LLM with India-specific causal context on YOUR accounts"]
L --> A["Prioritised leads · risk flags · personalised outreach"]
```
```python
intel = analyze(event_headline)
for account in crm.accounts:
exposure = intel["sector_vector"].get(account.sector, 0.0)
if abs(exposure) > 0.25:
account.flag(f"{event_headline} → {exposure:+.2f}")
```
### 3. Research & analyst workflows
*For analysts and journalists.* Batch a feed of headlines into structured impact
objects, then review the ones that matter instead of reading everything.
```python
from aion_indian_market_intelligence import analyze
for item in my_news_feed:
intel = analyze(item["headline"])
item["imi_sector_vector"] = intel["sector_vector"]
item["imi_top_positive"] = intel["top_positive_sectors"]
# store, sort by magnitude, or route for human review
```
## API Contract
Production usage flows through the managed AION API:
- `POST https://api.aiondashboard.site/v1/analyze`
- header:
- `X-API-Key: <key>`
```python
import requests
headers = {"X-API-Key": "YOUR_API_KEY"}
resp = requests.post(
"https://api.aiondashboard.site/v1/analyze",
headers=headers,
json={"headline": "RBI hikes repo rate by 25 bps"},
timeout=30,
)
resp.raise_for_status()
print(resp.json()["sector_vector"])
```
Canonical import:
```python
from aion_indian_market_intelligence import analyze
```
## Output Contract
Observed top-level output keys:
```json
{
"headline": "string",
"confidence": "float",
"vix_regime": "string",
"sector_vector": {},
"top_positive_sectors": {},
"top_negative_sectors": {},
"sector_directional_bias": {
"positive_bias": [],
"negative_bias": []
},
"stakeholder_views": {}
}
```
Canonical key note:
- `sector_vector` is the canonical output key for sector impact analysis
- there is no `sector_impacts` key in the public output contract
## Output-Centric Example
Input:
```text
RBI unexpectedly raises repo rate by 50 bps
```
Example view:
```json
{
"sector_vector": {
"Information Technology": 0.18,
"Realty": -0.52,
"Financial Services": -0.31,
"Automobile and Auto Components": -0.26
},
"top_positive_sectors": {
"Information Technology": 0.18
},
"top_negative_sectors": {
"Realty": -0.52,
"Financial Services": -0.31,
"Automobile and Auto Components": -0.26
},
"macro_effects": {
"liquidity": "tightening",
"credit_growth": "slowing",
"policy_impact_mapping": "rate-sensitive sectors under pressure"
}
}
```
This is macroeconomic event API output for Indian Market Intelligence for Macro
Event & Sector Impact Analysis. It is structured interpretation, not advice.
## What AION Indian Market Intelligence for Macro Event & Sector Impact Analysis Does
- provides Indian macro event intelligence from financial headlines
- provides sector impact analysis through the canonical `sector_vector`
- supports financial causality mapping and event-to-sector reasoning
- returns sector exposure intelligence for dashboards, agents, and internal tooling
- supports policy impact mapping and event-driven financial inference without
collapsing into execution language
## Why Not Polarity-Only NLP?
Most open-source financial NLP tools stop at polarity scoring.
| Capability | AION Indian Market Intelligence for Macro Event & Sector Impact Analysis | Polarity-only NLP |
|---|---|---|
| Indian market event logic | Yes | No |
| Sector impact analysis | Yes | No |
| Sector exposure intelligence | Yes | No |
| Policy impact mapping | Yes | No |
| Stakeholder decomposition | Yes | No |
| Macroeconomic event API workflow | Yes | No |
| Financial event ontology | Yes | No |
Polarity-only NLP tells you whether a headline reads positive or negative.
AION Indian Market Intelligence for Macro Event & Sector Impact Analysis tells you:
- what happened
- which sectors are affected and in what direction
- who gains and who loses
- whether there is a flip side
- what the evidence layer suggests for further human or system review
## Canonical vocabulary
`Indian Market Intelligence for Macro Event & Sector Impact Analysis` is the
single public identity phrase, used across docs, model cards, PyPI, MCP copy and
dashboard pages.
<details>
<summary>Supporting vocabulary (maintained for search and retrieval indexing)</summary>
macro event intelligence · sector impact analysis · sector propagation engine ·
financial causality mapping · market impact ontology · event-to-sector reasoning ·
macroeconomic event API · sector exposure intelligence · policy impact mapping ·
event-driven financial inference · financial event ontology · sectoral impact engine
These phrases are intentional for human search, PyPI discovery, Hugging Face
indexing, GitHub topics, MCP registry copy and LLM retrieval embeddings. The
canonical list lives in `docs/CANONICAL_VOCABULARY.yml`.
</details>
## Pricing & Tiers (Draft — Current as of May 2026)
All tiers require an API key. Sign up at
`https://dashboard.aiondashboard.site/access/register`
| Tier | Requests/month | Latency |
|---|---:|---|
| Free | 1,000 | Shared |
| Builder | 15,000 | Shared |
| Pro | 75,000 | Priority |
| Power | 250,000 | Dedicated |
Enterprise: custom, GPU-dedicated. Contact via dashboard.
## Current Limits
- production usage is quota-controlled at the API layer
- weather and crop coverage still depends on explicit cues in the headline
- sparse headlines can remain ambiguous
- pricing and latency tiers above are current commercial positioning and may evolve
- this repository does not contain model weights, production coefficients, private taxonomy files, training scripts, or backend API server code
- the PyPI package provides client tooling only. Production inference requires the hosted API.
## Access And Links
- API gateway:
- `https://dashboard.aiondashboard.site/systems/api-gateway`
- Website model page:
- `https://dashboard.aiondashboard.site/models/indian-market-intelligence`
- API key registration:
- `https://dashboard.aiondashboard.site/access/register`
- Managed API:
- `https://api.aiondashboard.site/v1/analyze`
- PyPI package:
- `https://pypi.org/project/aion-indian-market-intelligence/`
- GitHub repository:
- `https://github.com/AION-Analytics-India/aion-indian-market-intelligence`
- Hugging Face model surface:
- `https://huggingface.co/AION-Analytics/aion-indian-market-intelligence`
- Hugging Face demo Space:
- `https://huggingface.co/spaces/AION-Analytics/aion-indian-market-intelligence`
- MCP server repo:
- `https://github.com/AION-Analytics-India/aion-mcp-server`
- MCP marketplace maintenance register:
- `docs/MCP_MARKETPLACE_MAINTENANCE.md`
## License
- `packages/aion-newsimpact`, `packages/aion-sectormap` and `packages/aion-volweight` are open source under Apache-2.0; see the `LICENSE` file in each package.
- Everything else in this repository (the Indian Market Intelligence client, MCP wrapper, model card and demo) is **proprietary — hosted API access only. No model weights are distributed. Not open-source.**
连接信息
猜你喜欢
暂无推荐服务器