@wentorai/research-plugins 1.1.0 → 1.2.0

package/skills/domains/finance/finsight-research-guide/SKILL.md

@@ -0,0 +1,113 @@
+---
+name: finsight-research-guide
+description: "Deep financial research with the FinSight multi-agent system"
+metadata:
+  openclaw:
+    emoji: "💰"
+    category: "domains"
+    subcategory: "finance"
+    keywords: ["FinSight", "financial analysis", "deep research", "market analysis", "financial reports", "multi-agent"]
+    source: "https://github.com/RUC-NLPIR/FinSight"
+---
+
+# FinSight Research Guide
+
+## Overview
+
+FinSight is a deep research agent designed specifically for financial analysis. Developed by RUC-NLPIR, it combines multi-source data retrieval, financial reasoning, and report generation to produce publication-ready financial research. It handles market analysis, company fundamentals, sector comparisons, and macroeconomic assessment through specialized agents.
+
+## Installation
+
+```bash
+git clone https://github.com/RUC-NLPIR/FinSight.git
+cd FinSight && pip install -e .
+```
+
+## Core Capabilities
+
+### Research Query to Report
+
+```python
+from finsight import FinSightAgent
+
+agent = FinSightAgent(llm_provider="anthropic")
+
+# Generate comprehensive financial analysis
+report = agent.research(
+    "Analyze the competitive landscape of the global EV battery "
+    "market. Compare CATL, LG Energy, and Panasonic on market "
+    "share, technology, margins, and growth outlook."
+)
+
+print(report.summary)
+report.save("ev_battery_analysis.pdf")
+```
+
+### Agent Architecture
+
+| Agent | Role |
+|-------|------|
+| **Retrieval Agent** | Fetches data from SEC filings, financial APIs, news |
+| **Data Agent** | Processes financial statements, ratios, time series |
+| **Analysis Agent** | Performs fundamental, technical, and comparative analysis |
+| **Reasoning Agent** | Synthesizes findings, identifies trends and risks |
+| **Report Agent** | Generates structured research reports with citations |
+
+### Financial Data Sources
+
+```python
+# FinSight integrates with multiple data sources
+config = {
+    "sec_edgar": True,        # SEC filings (free)
+    "fred": True,             # Federal Reserve economic data
+    "yahoo_finance": True,    # Market data (free)
+    "news_api": True,         # Financial news
+    "world_bank": True,       # Macro indicators
+}
+```
+
+### Analysis Types
+
+```python
+# Company fundamental analysis
+report = agent.research(
+    "Provide a fundamental analysis of NVIDIA including "
+    "revenue trends, margin analysis, valuation multiples, "
+    "and competitive moat assessment."
+)
+
+# Sector analysis
+report = agent.research(
+    "Compare the top 5 cloud computing companies by revenue "
+    "growth, operating margins, and R&D investment intensity."
+)
+
+# Macro analysis
+report = agent.research(
+    "Analyze the impact of rising interest rates on US "
+    "commercial real estate valuations since 2022."
+)
+```
+
+## Report Structure
+
+Generated reports typically include:
+
+1. **Executive Summary** — Key findings in 3-5 bullets
+2. **Market Overview** — Industry size, growth, trends
+3. **Company Analysis** — Financials, competitive position
+4. **Risk Assessment** — Key risks and mitigation
+5. **Outlook** — Forward-looking analysis with scenarios
+6. **Sources** — Cited data sources and references
+
+## Use Cases
+
+1. **Investment research**: Company and sector deep dives
+2. **Due diligence**: Comprehensive target company analysis
+3. **Academic research**: Financial economics research support
+4. **Market intelligence**: Competitive landscape mapping
+
+## References
+
+- [FinSight GitHub](https://github.com/RUC-NLPIR/FinSight)
+- [RUC-NLPIR Lab](http://playbigdata.ruc.edu.cn/)

package/skills/domains/finance/options-analytics-agent-guide/SKILL.md

@@ -0,0 +1,117 @@
+---
+name: options-analytics-agent-guide
+description: "AI agent for options pricing, Greeks, and strategy analysis"
+metadata:
+  openclaw:
+    emoji: "📉"
+    category: "domains"
+    subcategory: "finance"
+    keywords: ["options analytics", "derivatives", "Greeks", "Black-Scholes", "strategy analysis", "financial agent"]
+    source: "https://github.com/options-analytics/options-agent"
+---
+
+# Options Analytics Agent Guide
+
+## Overview
+
+An AI agent for options pricing, risk analysis, and strategy evaluation. It combines Black-Scholes and binomial models, Greeks calculations, implied volatility surfaces, and portfolio risk analytics into a conversational interface. Researchers and quantitative analysts can query options data, price exotic derivatives, and evaluate trading strategies through natural language.
+
+## Core Capabilities
+
+```python
+from options_agent import OptionsAgent
+
+agent = OptionsAgent(llm_provider="anthropic")
+
+# Price an option
+result = agent.price(
+    option_type="call",
+    strike=100,
+    spot=105,
+    expiry_days=30,
+    risk_free_rate=0.05,
+    volatility=0.20,
+    model="black_scholes",
+)
+
+print(f"Price: ${result.price:.2f}")
+print(f"Delta: {result.delta:.4f}")
+print(f"Gamma: {result.gamma:.4f}")
+print(f"Theta: {result.theta:.4f}")
+print(f"Vega: {result.vega:.4f}")
+print(f"Rho: {result.rho:.4f}")
+```
+
+## Greeks Analysis
+
+```python
+# Full Greeks surface
+surface = agent.greeks_surface(
+    strike=100,
+    spot_range=(80, 120),
+    expiry_range=(7, 90),  # days
+    volatility=0.25,
+)
+
+surface.plot_delta_surface("delta_surface.png")
+surface.plot_gamma_surface("gamma_surface.png")
+surface.plot_theta_decay("theta_decay.png")
+```
+
+## Strategy Evaluation
+
+```python
+# Evaluate an options strategy
+strategy = agent.evaluate_strategy(
+    legs=[
+        {"type": "call", "strike": 100, "action": "buy", "qty": 1},
+        {"type": "call", "strike": 110, "action": "sell", "qty": 1},
+    ],
+    spot=105,
+    expiry_days=30,
+    volatility=0.20,
+)
+
+print(f"Strategy: {strategy.name}")  # Bull Call Spread
+print(f"Max profit: ${strategy.max_profit:.2f}")
+print(f"Max loss: ${strategy.max_loss:.2f}")
+print(f"Breakeven: ${strategy.breakeven:.2f}")
+
+strategy.plot_payoff("payoff.png")
+strategy.plot_pnl_scenarios("scenarios.png")
+```
+
+## Implied Volatility
+
+```python
+# Calculate implied volatility
+iv = agent.implied_volatility(
+    market_price=5.50,
+    option_type="call",
+    strike=100,
+    spot=105,
+    expiry_days=30,
+    risk_free_rate=0.05,
+)
+print(f"Implied volatility: {iv:.2%}")
+
+# Volatility smile/surface
+vol_surface = agent.volatility_surface(
+    ticker="SPY",
+    date="2025-03-10",
+)
+vol_surface.plot("vol_surface.png")
+```
+
+## Use Cases
+
+1. **Options pricing**: Black-Scholes and numerical methods
+2. **Risk management**: Greeks and portfolio risk metrics
+3. **Strategy analysis**: P&L profiles and breakeven analysis
+4. **Volatility analysis**: IV surfaces and skew analysis
+5. **Education**: Interactive derivatives teaching tool
+
+## References
+
+- [Options Analytics Agent](https://github.com/options-analytics/options-agent)
+- [QuantLib](https://www.quantlib.org/) — Quantitative finance library

package/skills/domains/geoscience/pangaea-data-api/SKILL.md

@@ -0,0 +1,197 @@
+---
+name: pangaea-data-api
+description: "Access earth and environmental science datasets via PANGAEA API"
+metadata:
+  openclaw:
+    emoji: "🌍"
+    category: "domains"
+    subcategory: "geoscience"
+    keywords: ["PANGAEA", "earth science data", "oceanography", "paleoclimate", "environmental data", "geoscience repository"]
+    source: "https://www.pangaea.de/"
+---
+
+# PANGAEA Data Repository API
+
+## Overview
+
+PANGAEA is the world's leading data repository for earth and environmental sciences, hosting 400K+ datasets with 20B+ data points. It archives research data from oceanography, paleoclimatology, geology, ecology, and atmospheric science. Each dataset has a DOI and is linked to the originating publication. The API provides search, metadata retrieval, and data download. Free, no authentication required.
+
+## API Endpoints
+
+### Search API
+
+```bash
+# Search datasets by keyword
+curl "https://www.pangaea.de/advanced/search.php?q=ocean+temperature&count=20&type=json"
+
+# Search with geographic bounding box
+curl "https://www.pangaea.de/advanced/search.php?\
+q=sediment+core&minlat=-60&maxlat=-30&minlon=-180&maxlon=180&type=json"
+
+# Filter by parameter (measurement type)
+curl "https://www.pangaea.de/advanced/search.php?\
+q=carbon+dioxide&param=Atmospheric+CO2&type=json"
+
+# Filter by date range
+curl "https://www.pangaea.de/advanced/search.php?\
+q=Arctic+ice&mindate=2020-01-01&maxdate=2026-12-31&type=json"
+```
+
+### ElasticSearch API
+
+```bash
+# Full-text search via Elasticsearch
+curl -X POST "https://ws.pangaea.de/es/pangaea/panmd/_search" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "query": {
+      "bool": {
+        "must": [
+          {"match": {"citation.title": "ocean temperature"}}
+        ],
+        "filter": [
+          {"range": {"citation.year": {"gte": 2020}}}
+        ]
+      }
+    },
+    "size": 20
+  }'
+```
+
+### Dataset Access
+
+```bash
+# Get dataset metadata
+curl "https://doi.pangaea.de/10.1594/PANGAEA.123456?format=metainfo_json"
+
+# Download dataset as tab-delimited text
+curl "https://doi.pangaea.de/10.1594/PANGAEA.123456?format=textfile"
+
+# Download as CSV
+curl "https://doi.pangaea.de/10.1594/PANGAEA.123456?format=csv"
+```
+
+### OAI-PMH Harvesting
+
+```bash
+# List records
+curl "https://ws.pangaea.de/oai/provider?verb=ListRecords&metadataPrefix=oai_dc"
+
+# Get specific record
+curl "https://ws.pangaea.de/oai/provider?verb=GetRecord&identifier=oai:pangaea.de:doi:10.1594/PANGAEA.123456&metadataPrefix=oai_dc"
+```
+
+### Query Parameters (Search API)
+
+| Parameter | Description | Example |
+|-----------|-------------|---------|
+| `q` | Search query | `q=coral+reef+bleaching` |
+| `count` | Results per page | `count=50` |
+| `offset` | Pagination offset | `offset=20` |
+| `minlat/maxlat` | Latitude bounds | `-90` to `90` |
+| `minlon/maxlon` | Longitude bounds | `-180` to `180` |
+| `mindate/maxdate` | Temporal filter | `2020-01-01` |
+| `param` | Parameter/measurement | `Temperature` |
+| `topic` | Topic filter | `Atmosphere`, `Biosphere` |
+| `type` | Response format | `json`, `xml` |
+
+## Python Usage
+
+```python
+import requests
+import pandas as pd
+from io import StringIO
+
+SEARCH_URL = "https://www.pangaea.de/advanced/search.php"
+ES_URL = "https://ws.pangaea.de/es/pangaea/panmd/_search"
+
+
+def search_pangaea(query: str, count: int = 20,
+                   bbox: dict = None) -> list:
+    """Search PANGAEA for earth science datasets."""
+    params = {"q": query, "count": count, "type": "json"}
+    if bbox:
+        params.update({
+            "minlat": bbox.get("south", -90),
+            "maxlat": bbox.get("north", 90),
+            "minlon": bbox.get("west", -180),
+            "maxlon": bbox.get("east", 180),
+        })
+
+    resp = requests.get(SEARCH_URL, params=params, timeout=30)
+    resp.raise_for_status()
+    data = resp.json()
+
+    results = []
+    for item in data.get("results", []):
+        results.append({
+            "doi": item.get("URI", ""),
+            "title": item.get("citation", ""),
+            "year": item.get("year"),
+            "size": item.get("size"),
+            "parameters": item.get("params", []),
+            "score": item.get("score"),
+        })
+    return results
+
+
+def download_dataset(doi: str) -> pd.DataFrame:
+    """Download a PANGAEA dataset as a pandas DataFrame."""
+    url = f"https://doi.pangaea.de/{doi}?format=textfile"
+    resp = requests.get(url, timeout=60)
+    resp.raise_for_status()
+
+    lines = resp.text.split("\n")
+    header_end = next(
+        (i for i, line in enumerate(lines) if line.startswith("*/")),
+        -1,
+    )
+    data_text = "\n".join(lines[header_end + 1:])
+    return pd.read_csv(StringIO(data_text), sep="\t")
+
+
+def search_by_location(query: str, lat: float, lon: float,
+                       radius_deg: float = 5.0) -> list:
+    """Search datasets near a geographic location."""
+    bbox = {
+        "south": lat - radius_deg,
+        "north": lat + radius_deg,
+        "west": lon - radius_deg,
+        "east": lon + radius_deg,
+    }
+    return search_pangaea(query, bbox=bbox)
+
+
+# Example: find ocean temperature datasets
+datasets = search_pangaea("sea surface temperature", count=5)
+for ds in datasets:
+    print(f"[{ds['year']}] {ds['title'][:80]}...")
+    print(f"  DOI: {ds['doi']} | Size: {ds['size']}")
+
+# Example: download a specific dataset
+# df = download_dataset("10.1594/PANGAEA.123456")
+# print(df.head())
+
+# Example: find Arctic research data
+arctic = search_by_location("permafrost", lat=70, lon=25)
+for ds in arctic[:3]:
+    print(f"{ds['title'][:80]}...")
+```
+
+## Data Topics
+
+| Topic | Coverage |
+|-------|----------|
+| Oceans | Temperature, salinity, currents, chemistry |
+| Paleoclimate | Ice cores, sediment cores, tree rings |
+| Atmosphere | CO2, aerosols, weather observations |
+| Lithosphere | Geology, tectonics, geochemistry |
+| Biosphere | Biodiversity, ecology, marine biology |
+| Cryosphere | Sea ice, glaciers, permafrost |
+
+## References
+
+- [PANGAEA](https://www.pangaea.de/)
+- [PANGAEA API](https://wiki.pangaea.de/wiki/PANGAEA_API)
+- [PANGAEA Elasticsearch](https://wiki.pangaea.de/wiki/Elasticsearch_API)
+- Diepenbroek, M. et al. (2002). "PANGAEA — an information system for environmental sciences." *C&G* 28(10).

package/skills/domains/humanities/digital-humanities-methods/SKILL.md

@@ -0,0 +1,232 @@
+---
+name: digital-humanities-methods
+description: "Computational methods for humanities research with text and network analysis"
+metadata:
+  openclaw:
+    emoji: "📜"
+    category: "domains"
+    subcategory: "humanities"
+    keywords: ["digital humanities", "text analysis", "corpus linguistics", "network analysis", "cultural analytics", "computational methods"]
+    source: "https://clawhub.ai/digital-humanities"
+---
+
+# Digital Humanities Methods
+
+## Overview
+
+Digital Humanities (DH) applies computational methods to humanistic inquiry — analyzing literary texts, historical records, cultural artifacts, and social networks at scale. This guide covers the core computational methods used in DH research: text analysis, topic modeling, network analysis, spatial analysis, and corpus linguistics. These methods complement rather than replace traditional close reading and archival research.
+
+## Text Analysis
+
+### Preprocessing Pipeline
+
+```python
+import re
+from collections import Counter
+
+def preprocess_text(text: str, language: str = "en") -> list:
+    """Standard preprocessing for humanities text analysis."""
+    # Lowercase
+    text = text.lower()
+
+    # Remove metadata markers (page numbers, headers)
+    text = re.sub(r'\[page \d+\]', '', text)
+    text = re.sub(r'\n{3,}', '\n\n', text)
+
+    # Tokenize (simple whitespace + punctuation split)
+    tokens = re.findall(r'\b[a-z]+\b', text)
+
+    # Remove stopwords (customize per corpus!)
+    # Standard lists often remove words meaningful in literary analysis
+    # e.g., "not", "but", "never" carry sentiment — keep them if relevant
+    stopwords = {"the", "a", "an", "is", "are", "was", "were", "in",
+                 "on", "at", "to", "for", "of", "with", "and", "or"}
+    tokens = [t for t in tokens if t not in stopwords and len(t) > 2]
+
+    return tokens
+
+# Word frequency analysis
+def word_frequencies(tokens: list, top_n: int = 50) -> list:
+    return Counter(tokens).most_common(top_n)
+```
+
+### Stylometry (Authorship Analysis)
+
+```python
+"""Stylometric features for authorship attribution."""
+
+def extract_style_features(text: str) -> dict:
+    """Extract stylistic features from a text."""
+    sentences = text.split('.')
+    words = text.split()
+    chars = list(text)
+
+    return {
+        "avg_sentence_length": len(words) / max(len(sentences), 1),
+        "avg_word_length": sum(len(w) for w in words) / max(len(words), 1),
+        "vocabulary_richness": len(set(words)) / max(len(words), 1),
+        "hapax_ratio": sum(1 for w, c in Counter(words).items() if c == 1) / max(len(set(words)), 1),
+        "comma_rate": text.count(',') / max(len(words), 1),
+        "semicolon_rate": text.count(';') / max(len(words), 1),
+        "question_rate": text.count('?') / max(len(sentences), 1),
+        "exclamation_rate": text.count('!') / max(len(sentences), 1),
+    }
+```
+
+### Sentiment Analysis for Historical Texts
+
+```python
+# Note: Modern NLP sentiment tools are trained on contemporary text.
+# For historical texts, consider:
+# 1. Historical sentiment lexicons (e.g., NRC Emotion Lexicon adapted)
+# 2. Period-specific word lists
+# 3. Manual validation on a sample before scaling
+
+from transformers import pipeline
+
+# Modern text sentiment (use with caution on historical texts)
+sentiment = pipeline("sentiment-analysis")
+result = sentiment("It was the best of times, it was the worst of times.")
+
+# Better: keyword-based approach with custom lexicons
+POSITIVE = {"virtue", "honor", "glory", "triumph", "beauty", "noble"}
+NEGATIVE = {"vice", "shame", "defeat", "ruin", "wretched", "base"}
+
+def lexicon_sentiment(tokens: list, pos: set, neg: set) -> float:
+    """Simple lexicon-based sentiment score."""
+    pos_count = sum(1 for t in tokens if t in pos)
+    neg_count = sum(1 for t in tokens if t in neg)
+    total = pos_count + neg_count
+    if total == 0:
+        return 0.0
+    return (pos_count - neg_count) / total
+```
+
+## Topic Modeling
+
+```python
+from sklearn.feature_extraction.text import CountVectorizer
+from sklearn.decomposition import LatentDirichletAllocation
+
+def run_topic_model(documents: list, n_topics: int = 10,
+                     n_top_words: int = 15):
+    """LDA topic modeling on a corpus of documents."""
+    # Vectorize
+    vectorizer = CountVectorizer(max_df=0.95, min_df=2,
+                                  max_features=5000,
+                                  stop_words="english")
+    dtm = vectorizer.fit_transform(documents)
+    feature_names = vectorizer.get_feature_names_out()
+
+    # Fit LDA
+    lda = LatentDirichletAllocation(n_components=n_topics,
+                                     random_state=42,
+                                     max_iter=50)
+    lda.fit(dtm)
+
+    # Display topics
+    topics = {}
+    for topic_idx, topic in enumerate(lda.components_):
+        top_words = [feature_names[i] for i in topic.argsort()[-n_top_words:]]
+        topics[f"Topic {topic_idx}"] = top_words
+        print(f"Topic {topic_idx}: {', '.join(top_words)}")
+
+    return lda, topics, dtm
+```
+
+## Network Analysis
+
+### Social Network from Letters/Correspondence
+
+```python
+import networkx as nx
+
+def build_correspondence_network(letters: list) -> nx.Graph:
+    """Build a social network from letter metadata.
+
+    Args:
+        letters: list of dicts with 'sender', 'recipient', 'date', 'location'
+    """
+    G = nx.Graph()
+
+    for letter in letters:
+        sender = letter["sender"]
+        recipient = letter["recipient"]
+
+        if G.has_edge(sender, recipient):
+            G[sender][recipient]["weight"] += 1
+        else:
+            G.add_edge(sender, recipient, weight=1)
+
+    # Compute centrality measures
+    centrality = nx.degree_centrality(G)
+    betweenness = nx.betweenness_centrality(G)
+
+    print(f"Network: {G.number_of_nodes()} individuals, "
+          f"{G.number_of_edges()} connections")
+    print(f"Most connected: {max(centrality, key=centrality.get)}")
+    print(f"Most bridging: {max(betweenness, key=betweenness.get)}")
+
+    return G
+```
+
+## Spatial Analysis (GIS for Humanities)
+
+Common applications:
+- Mapping historical events or migration patterns
+- Georeferencing historical maps
+- Spatial analysis of literary settings
+
+```python
+import folium
+
+def create_historical_map(events: list, title: str = "Historical Events"):
+    """Create an interactive map of historical events.
+
+    Args:
+        events: list of dicts with 'name', 'lat', 'lon', 'date', 'description'
+    """
+    center_lat = sum(e["lat"] for e in events) / len(events)
+    center_lon = sum(e["lon"] for e in events) / len(events)
+
+    m = folium.Map(location=[center_lat, center_lon], zoom_start=5)
+
+    for event in events:
+        popup = f"<b>{event['name']}</b><br>{event['date']}<br>{event['description']}"
+        folium.Marker(
+            location=[event["lat"], event["lon"]],
+            popup=popup,
+            tooltip=event["name"]
+        ).add_to(m)
+
+    m.save(f"{title.replace(' ', '_').lower()}.html")
+    return m
+```
+
+## Key Data Sources for DH
+
+| Source | Content | Access |
+|--------|---------|--------|
+| **Project Gutenberg** | 70,000+ free ebooks | gutenberg.org |
+| **HathiTrust** | 17M+ digitized volumes | hathitrust.org |
+| **Internet Archive** | Books, media, web archives | archive.org |
+| **EEBO / ECCO** | Early English books (1475-1800) | Institutional |
+| **Perseus Digital Library** | Greek and Latin classics | perseus.tufts.edu |
+| **Europeana** | European cultural heritage | europeana.eu |
+| **DPLA** | US digital public library | dp.la |
+| **Old Bailey Online** | London criminal trials (1674-1913) | oldbaileyonline.org |
+
+## Methodological Considerations
+
+1. **Close reading still matters**: Computational methods reveal patterns; interpretation requires humanistic expertise
+2. **Corpus bias**: Digitized collections over-represent certain periods, languages, and genres
+3. **OCR quality**: Historical texts often have high OCR error rates — validate before analysis
+4. **Anachronism**: Modern NLP tools may misinterpret historical language use
+5. **Interdisciplinary collaboration**: Best DH work pairs domain expertise with technical skills
+
+## References
+
+- Moretti, F. (2013). *Distant Reading*. Verso.
+- Jockers, M. L. (2013). *Macroanalysis: Digital Methods and Literary History*. University of Illinois Press.
+- [Programming Historian](https://programminghistorian.org/) — Free tutorials for DH methods
+- [DH Tools Directory](https://dirtdirectory.org/)