fin-infra 0.2.1__tar.gz → 0.2.2__tar.gz

{fin_infra-0.2.1 → fin_infra-0.2.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: fin-infra
-Version: 0.2.1
+Version: 0.2.2
 Summary: Financial infrastructure toolkit: banking connections, market data, credit, cashflows, and brokerage integrations
 License: MIT
 Keywords: finance,banking,plaid,brokerage,markets,credit,tax,cashflow,fintech,infra
@@ -47,26 +47,27 @@ Project-URL: Issues, https://github.com/nfraxlab/fin-infra/issues
 Project-URL: Repository, https://github.com/nfraxlab/fin-infra
 Description-Content-Type: text/markdown
 
-<div align="center">
-
 # fin-infra
 
-[![CI](https://github.com/nfraxlab/fin-infra/actions/workflows/ci.yml/badge.svg)](https://github.com/nfraxlab/fin-infra/actions/workflows/ci.yml)
+**Financial data infrastructure for fintech apps.**
+
 [![PyPI](https://img.shields.io/pypi/v/fin-infra.svg)](https://pypi.org/project/fin-infra/)
+[![CI](https://github.com/nfraxlab/fin-infra/actions/workflows/ci.yml/badge.svg)](https://github.com/nfraxlab/fin-infra/actions/workflows/ci.yml)
 [![Python](https://img.shields.io/pypi/pyversions/fin-infra.svg)](https://pypi.org/project/fin-infra/)
 [![License](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE)
-[![Downloads](https://img.shields.io/pypi/dm/fin-infra.svg)](https://pypi.org/project/fin-infra/)
-[![codecov](https://codecov.io/gh/nfraxlab/fin-infra/branch/main/graph/badge.svg)](https://codecov.io/gh/nfraxlab/fin-infra)
-
-### Financial data infrastructure for fintech apps
 
-**Banking, investments, market data, credit scores, and financial calculations in one toolkit.**
+## Overview
 
-[Documentation](docs/) · [Examples](examples/) · [PyPI](https://pypi.org/project/fin-infra/)
+Banking, investments, market data, credit scores, and financial calculations in one toolkit.
 
-</div>
+### Key Features
 
----
+- **Banking** - Plaid/Teller integration, accounts, transactions
+- **Investments** - Holdings, portfolio data, real P/L with cost basis
+- **Market Data** - Stocks, crypto, forex quotes and history
+- **Credit** - Credit scores and monitoring
+- **Analytics** - Cash flow, savings rate, spending insights
+- **Cashflows** - NPV, IRR, loan amortization calculations
 
 ## Why fin-infra?
 

{fin_infra-0.2.1 → fin_infra-0.2.2}/README.md

@@ -1,23 +1,24 @@
-<div align="center">
-
 # fin-infra
 
-[![CI](https://github.com/nfraxlab/fin-infra/actions/workflows/ci.yml/badge.svg)](https://github.com/nfraxlab/fin-infra/actions/workflows/ci.yml)
+**Financial data infrastructure for fintech apps.**
+
 [![PyPI](https://img.shields.io/pypi/v/fin-infra.svg)](https://pypi.org/project/fin-infra/)
+[![CI](https://github.com/nfraxlab/fin-infra/actions/workflows/ci.yml/badge.svg)](https://github.com/nfraxlab/fin-infra/actions/workflows/ci.yml)
 [![Python](https://img.shields.io/pypi/pyversions/fin-infra.svg)](https://pypi.org/project/fin-infra/)
 [![License](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE)
-[![Downloads](https://img.shields.io/pypi/dm/fin-infra.svg)](https://pypi.org/project/fin-infra/)
-[![codecov](https://codecov.io/gh/nfraxlab/fin-infra/branch/main/graph/badge.svg)](https://codecov.io/gh/nfraxlab/fin-infra)
-
-### Financial data infrastructure for fintech apps
 
-**Banking, investments, market data, credit scores, and financial calculations in one toolkit.**
+## Overview
 
-[Documentation](docs/) · [Examples](examples/) · [PyPI](https://pypi.org/project/fin-infra/)
+Banking, investments, market data, credit scores, and financial calculations in one toolkit.
 
-</div>
+### Key Features
 
----
+- **Banking** - Plaid/Teller integration, accounts, transactions
+- **Investments** - Holdings, portfolio data, real P/L with cost basis
+- **Market Data** - Stocks, crypto, forex quotes and history
+- **Credit** - Credit scores and monitoring
+- **Analytics** - Cash flow, savings rate, spending insights
+- **Cashflows** - NPV, IRR, loan amortization calculations
 
 ## Why fin-infra?
 

{fin_infra-0.2.1 → fin_infra-0.2.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "fin-infra"
-version = "0.2.1"
+version = "0.2.2"
 description = "Financial infrastructure toolkit: banking connections, market data, credit, cashflows, and brokerage integrations"
 authors = ["Ali Khatami <aliikhatami94@gmail.com>"]
 license = "MIT"

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/analytics/add.py

@@ -239,9 +239,9 @@ def add_analytics(
 
         Note:
             Real holdings provide:
-            - Accurate cost basis → real profit/loss
-            - Security types → precise asset allocation
-            - Current values → live portfolio tracking
+            - Accurate cost basis -> real profit/loss
+            - Security types -> precise asset allocation
+            - Current values -> live portfolio tracking
         """
         # If with_holdings requested and investment provider available
         if with_holdings:

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/analytics/portfolio.py

@@ -527,10 +527,10 @@ def portfolio_metrics_with_holdings(holdings: list) -> PortfolioMetrics:
         PortfolioMetrics with real portfolio analysis
 
     Real Data Advantages:
-        - Actual cost basis → accurate P/L calculations
-        - Real security types → precise asset allocation
-        - Current market values → live portfolio value
-        - No mock data → production-ready analytics
+        - Actual cost basis -> accurate P/L calculations
+        - Real security types -> precise asset allocation
+        - Current market values -> live portfolio value
+        - No mock data -> production-ready analytics
 
     Limitations:
         - Day/YTD/MTD returns require historical snapshots (not in holdings)
@@ -652,7 +652,7 @@ def calculate_day_change_with_snapshot(
         Function matches holdings by account_id + security_id for accurate tracking
         of individual position changes (accounts for buys/sells, not just price moves).
     """
-    # Build lookup map for previous snapshot: (account_id, security_id) → value
+    # Build lookup map for previous snapshot: (account_id, security_id) -> value
     previous_map = {}
     for holding in previous_snapshot:
         key = (holding.account_id, holding.security.security_id)
@@ -703,13 +703,13 @@ def _calculate_allocation_from_holdings(
         list[AssetAllocation] with asset_class, value, and percentage
 
     Asset Class Mapping:
-        - equity → Stocks
-        - etf → Stocks (equity ETFs grouped with stocks)
-        - mutual_fund → Bonds (conservative assumption)
-        - bond → Bonds
-        - cash → Cash
-        - derivative → Other
-        - other → Other
+        - equity -> Stocks
+        - etf -> Stocks (equity ETFs grouped with stocks)
+        - mutual_fund -> Bonds (conservative assumption)
+        - bond -> Bonds
+        - cash -> Cash
+        - derivative -> Other
+        - other -> Other
     """
     from collections import defaultdict
 

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/brokerage/__init__.py

@@ -54,8 +54,8 @@ def easy_brokerage(
     [!] **SAFETY**: Defaults to paper trading mode. Live trading requires explicit mode="live".
 
     Auto-detects provider based on environment variables:
-    1. If ALPACA_API_KEY and ALPACA_API_SECRET are set → Alpaca
-    2. Otherwise → Raises error (credentials required)
+    1. If ALPACA_API_KEY and ALPACA_API_SECRET are set -> Alpaca
+    2. Otherwise -> Raises error (credentials required)
 
     Args:
         provider: Provider name ("alpaca"). If None, defaults to alpaca.

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/categorization/__init__.py

@@ -2,7 +2,7 @@
 Transaction categorization module.
 
 Provides ML-based categorization of merchant transactions into 56 categories
-using a hybrid approach (exact match → regex → sklearn Naive Bayes → LLM).
+using a hybrid approach (exact match -> regex -> sklearn Naive Bayes -> LLM).
 
 Basic usage:
     from fin_infra.categorization import categorize

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/categorization/engine.py

@@ -1,5 +1,5 @@
 """
-Hybrid categorization engine (exact → regex → ML → LLM).
+Hybrid categorization engine (exact -> regex -> ML -> LLM).
 
 4-layer approach:
 1. Layer 1 (Exact Match): O(1) dictionary lookup, 85-90% coverage

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/categorization/llm_layer.py

@@ -197,7 +197,7 @@ class LLMCategorizer:
             self._track_cost()
 
             logger.info(
-                f"LLM categorized '{merchant_name}' → {prediction.category} "
+                f"LLM categorized '{merchant_name}' -> {prediction.category} "
                 f"(confidence={prediction.confidence:.2f})"
             )
 
@@ -252,7 +252,7 @@ class LLMCategorizer:
         # Format few-shot examples
         examples_text = "\n\n".join(
             [
-                f'Merchant: "{merchant}"\n→ Category: "{category}"\n→ Reasoning: "{reasoning}"'
+                f'Merchant: "{merchant}"\n-> Category: "{category}"\n-> Reasoning: "{reasoning}"'
                 for merchant, category, reasoning in FEW_SHOT_EXAMPLES
             ]
         )
@@ -338,10 +338,10 @@ Return JSON with category, confidence, and reasoning."""
 
     def reset_daily_cost(self):
         """Reset daily cost counter (called at midnight UTC)."""
-        logger.info(f"Resetting daily cost: ${self.daily_cost:.5f} → $0.00")
+        logger.info(f"Resetting daily cost: ${self.daily_cost:.5f} -> $0.00")
         self.daily_cost = 0.0
 
     def reset_monthly_cost(self):
         """Reset monthly cost counter (called on 1st of month)."""
-        logger.info(f"Resetting monthly cost: ${self.monthly_cost:.5f} → $0.00")
+        logger.info(f"Resetting monthly cost: ${self.monthly_cost:.5f} -> $0.00")
         self.monthly_cost = 0.0

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/credit/experian/parser.py

@@ -1,11 +1,11 @@
 """Response parsers for Experian API data to fin_infra models.
 
 Converts Experian API JSON responses to typed Pydantic models:
-- parse_credit_score(): dict → CreditScore
-- parse_credit_report(): dict → CreditReport
-- parse_account(): dict → CreditAccount
-- parse_inquiry(): dict → CreditInquiry
-- parse_public_record(): dict → PublicRecord
+- parse_credit_score(): dict -> CreditScore
+- parse_credit_report(): dict -> CreditReport
+- parse_account(): dict -> CreditAccount
+- parse_inquiry(): dict -> CreditInquiry
+- parse_public_record(): dict -> PublicRecord
 
 Example:
     >>> data = await client.get_credit_score("user123")

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/crypto/__init__.py

@@ -29,8 +29,8 @@ def easy_crypto(
     """Create a crypto data provider with zero or minimal configuration.
 
     Auto-detects provider based on environment variables:
-    1. If COINGECKO_API_KEY is set → CoinGecko Pro
-    2. Otherwise → CoinGecko Free (no key needed)
+    1. If COINGECKO_API_KEY is set -> CoinGecko Pro
+    2. Otherwise -> CoinGecko Free (no key needed)
 
     Args:
         provider: Provider name ("coingecko"). If None, defaults to coingecko.

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/goals/models.py

@@ -93,8 +93,8 @@ class FundingSource(BaseModel):
 
     Supports split allocation:
     - Multiple accounts can fund one goal (e.g., savings + checking)
-    - One account can fund multiple goals (e.g., savings → emergency + vacation)
-    - Allocation percentages must sum to ≤100% per account
+    - One account can fund multiple goals (e.g., savings -> emergency + vacation)
+    - Allocation percentages must sum to <=100% per account
     """
 
     goal_id: str = Field(..., description="Goal identifier")

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/investments/__init__.py

@@ -74,8 +74,8 @@ def easy_investments(
         InvestmentProvider instance for fetching holdings, transactions, securities.
 
     Environment detection order:
-        1. If PLAID_CLIENT_ID set → Plaid
-        2. If SNAPTRADE_CLIENT_ID set → SnapTrade
+        1. If PLAID_CLIENT_ID set -> Plaid
+        2. If SNAPTRADE_CLIENT_ID set -> SnapTrade
         3. Default: Plaid (most common)
 
     Examples:

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/investments/providers/base.py

@@ -236,10 +236,10 @@ class InvestmentProvider(ABC):
             Standardized SecurityType enum value
 
         Example mappings:
-            Plaid: "equity" → SecurityType.equity
-            Plaid: "mutual fund" → SecurityType.mutual_fund
-            SnapTrade: "cs" → SecurityType.equity (common stock)
-            SnapTrade: "etf" → SecurityType.etf
+            Plaid: "equity" -> SecurityType.equity
+            Plaid: "mutual fund" -> SecurityType.mutual_fund
+            SnapTrade: "cs" -> SecurityType.equity (common stock)
+            SnapTrade: "etf" -> SecurityType.etf
 
         Note:
             Override in provider-specific implementations for custom mappings.

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/markets/__init__.py

@@ -33,8 +33,8 @@ def easy_market(
     """Create a market data provider with zero or minimal configuration.
 
     Auto-detects provider based on environment variables:
-    1. If ALPHA_VANTAGE_API_KEY or ALPHAVANTAGE_API_KEY is set → Alpha Vantage
-    2. Otherwise → Yahoo Finance (no key needed)
+    1. If ALPHA_VANTAGE_API_KEY or ALPHAVANTAGE_API_KEY is set -> Alpha Vantage
+    2. Otherwise -> Yahoo Finance (no key needed)
 
     Args:
         provider: Provider name ("alphavantage" or "yahoo").

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/net_worth/__init__.py

@@ -13,7 +13,7 @@ Calculates net worth by aggregating balances from multiple financial providers
 
 **Key Features**:
 - Multi-provider aggregation (banking + brokerage + crypto)
-- Currency normalization (all currencies → USD)
+- Currency normalization (all currencies -> USD)
 - Historical snapshots (daily at midnight UTC)
 - Change detection (>5% or >$10k triggers webhook)
 - Asset allocation breakdown (pie charts)

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/net_worth/aggregator.py

@@ -59,7 +59,7 @@ class NetWorthAggregator:
     - Multi-provider support (banking, brokerage, crypto)
     - Parallel account fetching (faster performance)
     - Graceful error handling (continue if one provider fails)
-    - Currency normalization (all → base currency)
+    - Currency normalization (all -> base currency)
     - Market value calculation (stocks/crypto)
 
     **Example**:

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/net_worth/calculator.py

@@ -3,7 +3,7 @@ Net Worth Calculator Module
 
 Provides core calculation functions for net worth tracking:
 - Net worth calculation (assets - liabilities)
-- Currency normalization (all currencies → base currency)
+- Currency normalization (all currencies -> base currency)
 - Asset allocation breakdown
 - Change detection (amount + percentage)
 

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/net_worth/insights.py

@@ -167,7 +167,7 @@ Be specific with numbers. Cite percentage changes and dollar amounts.
 Focus on actionable insights, not generic advice.
 
 Example 1:
-User: Net worth: $500k → $575k over 6 months. Assets: +$65k (investments +$60k, savings +$5k). Liabilities: -$10k (new mortgage).
+User: Net worth: $500k -> $575k over 6 months. Assets: +$65k (investments +$60k, savings +$5k). Liabilities: -$10k (new mortgage).
 Response: {
   "summary": "Net worth increased 15% ($75k) over 6 months, driven primarily by strong investment performance.",
   "period": "6 months",
@@ -191,7 +191,7 @@ Response: {
 }
 
 Example 2:
-User: Net worth: $100k → $95k over 3 months. Assets: -$2k (market down). Liabilities: +$3k (credit card debt).
+User: Net worth: $100k -> $95k over 3 months. Assets: -$2k (market down). Liabilities: +$3k (credit card debt).
 Response: {
   "summary": "Net worth decreased 5% ($5k) over 3 months due to market decline and rising credit card debt.",
   "period": "3 months",
@@ -333,7 +333,7 @@ Given current allocation, age, and risk tolerance:
 3. Provide specific rebalancing steps
 
 Rule of thumb:
-- Stock allocation = 100 - age (e.g., age 35 → 65% stocks)
+- Stock allocation = 100 - age (e.g., age 35 -> 65% stocks)
 - Bonds for stability (increases with age)
 - Cash for emergency fund (3-6 months expenses)
 

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/normalization/__init__.py

@@ -54,7 +54,7 @@ def easy_normalization(
     Example:
         >>> from fin_infra.normalization import easy_normalization
         >>> resolver, converter = easy_normalization()
-        >>> ticker = await resolver.to_ticker("037833100")  # CUSIP → AAPL
+        >>> ticker = await resolver.to_ticker("037833100")  # CUSIP -> AAPL
         >>> eur = await converter.convert(100, "USD", "EUR")  # 92.0
     """
     global _resolver_instance, _converter_instance
@@ -107,7 +107,7 @@ def add_normalization(
         >>> resolver, converter = add_normalization(app)
         >>>
         >>> # Routes available:
-        >>> # GET /normalize/symbol/037833100 → {"ticker": "AAPL", ...}
+        >>> # GET /normalize/symbol/037833100 -> {"ticker": "AAPL", ...}
         >>> # GET /normalize/convert?amount=100&from_currency=USD&to_currency=EUR
 
     Integration with svc-infra:

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/normalization/currency_converter.py

@@ -71,7 +71,7 @@ class CurrencyConverter:
         except ExchangeRateAPIError as e:
             logger.error(f"Failed to convert {from_currency} to {to_currency}: {e}")
             raise CurrencyNotSupportedError(
-                f"Conversion failed: {from_currency} → {to_currency}"
+                f"Conversion failed: {from_currency} -> {to_currency}"
             ) from e
 
     async def get_rate(
@@ -108,9 +108,9 @@ class CurrencyConverter:
             return rate_data.rate
 
         except ExchangeRateAPIError as e:
-            logger.error(f"Failed to get rate {from_currency} → {to_currency}: {e}")
+            logger.error(f"Failed to get rate {from_currency} -> {to_currency}: {e}")
             raise CurrencyNotSupportedError(
-                f"Rate not available: {from_currency} → {to_currency}"
+                f"Rate not available: {from_currency} -> {to_currency}"
             ) from e
 
     async def get_rates(self, base_currency: str = "USD") -> dict[str, float]:

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/normalization/providers/static_mappings.py

@@ -109,7 +109,7 @@ TICKER_TO_ISIN = {
 }
 
 # Provider-specific symbol normalization
-# Maps provider-specific format → standard ticker
+# Maps provider-specific format -> standard ticker
 PROVIDER_SYMBOL_MAP = {
     "yahoo": {
         # Yahoo Finance uses dashes for crypto

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/obs/classifier.py

@@ -63,8 +63,8 @@ def financial_route_classifier(route_path: str, method: str) -> str:
     svc-infra's add_observability route_classifier parameter.
 
     Classification Logic:
-    - Financial routes (e.g., /banking/*, /market/*) → "financial"
-    - All other routes → "public"
+    - Financial routes (e.g., /banking/*, /market/*) -> "financial"
+    - All other routes -> "public"
 
     This allows Grafana dashboards to split metrics by route class:
     - Filter by route_class="financial" for financial provider SLOs

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/recurring/add.py

@@ -109,7 +109,7 @@ def add_recurring_detection(
         Detect recurring patterns in transaction history.
 
         Analyzes transaction history for recurring subscriptions and bills using
-        3-layer hybrid detection (fixed → variable → irregular).
+        3-layer hybrid detection (fixed -> variable -> irregular).
 
         **Example Request:**
         ```json

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/recurring/detectors_llm.py

@@ -100,30 +100,30 @@ Examples:
 1. Merchant: "City Electric"
    Amounts: [$45, $52, $48, $55, $50, $49]
    Dates: Monthly (15th ±7 days)
-   → is_recurring: true, cadence: "monthly", range: (40, 60),
+   -> is_recurring: true, cadence: "monthly", range: (40, 60),
      reasoning: "Seasonal winter heating variation", confidence: 0.85
 
 2. Merchant: "T-Mobile"
    Amounts: [$50, $50, $50, $78, $50, $50]
    Dates: Monthly (20th ±3 days)
-   → is_recurring: true, cadence: "monthly", range: (50, 80),
+   -> is_recurring: true, cadence: "monthly", range: (50, 80),
      reasoning: "Occasional overage charge spike", confidence: 0.80
 
 3. Merchant: "Random Store"
    Amounts: [$10, $45, $23, $67, $12]
    Dates: Irregular
-   → is_recurring: false, reasoning: "Too much variance, no pattern", confidence: 0.95
+   -> is_recurring: false, reasoning: "Too much variance, no pattern", confidence: 0.95
 
 4. Merchant: "Gas Company"
    Amounts: [$45, $48, $50, $52, $120, $115]
    Dates: Monthly
-   → is_recurring: true, cadence: "monthly", range: (40, 120),
+   -> is_recurring: true, cadence: "monthly", range: (40, 120),
      reasoning: "Winter heating season doubles bill", confidence: 0.80
 
 5. Merchant: "Gym Membership"
    Amounts: [$40, $40, $0, $40, $40]
    Dates: Monthly
-   → is_recurring: true, cadence: "monthly", range: (0, 40),
+   -> is_recurring: true, cadence: "monthly", range: (0, 40),
      reasoning: "Annual fee waived one month", confidence: 0.75
 
 Output format (JSON):

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/recurring/ease.py

@@ -43,7 +43,7 @@ def easy_recurring_detection(
     V2 Parameters (LLM Enhancement):
         enable_llm: Enable LLM for merchant normalization and variable detection (default: False)
                    When False, uses V1 pattern-based only (fast, $0 cost)
-                   When True, uses 4-layer hybrid (RapidFuzz → LLM normalization → Statistical → LLM variable detection)
+                   When True, uses 4-layer hybrid (RapidFuzz -> LLM normalization -> Statistical -> LLM variable detection)
         llm_provider: LLM provider to use (default: "google")
                      Options: "google" (Gemini 2.0 Flash, cheapest), "openai" (GPT-4o-mini), "anthropic" (Claude 3.5 Haiku)
         llm_model: Override default model for provider (default: None)
@@ -54,9 +54,9 @@ def easy_recurring_detection(
                                  Higher values (0.9) call LLM more often (more accurate, higher cost)
                                  Lower values (0.7) call LLM less often (less accurate, lower cost)
         llm_cache_merchant_ttl: Merchant normalization cache TTL in seconds (default: 604800 = 7 days)
-                               95% cache hit rate expected → most requests <1ms
+                               95% cache hit rate expected -> most requests <1ms
         llm_cache_insights_ttl: Insights generation cache TTL in seconds (default: 86400 = 24 hours)
-                               80% cache hit rate expected → most requests <1ms
+                               80% cache hit rate expected -> most requests <1ms
         llm_max_cost_per_day: Daily budget cap in USD (default: $0.10)
                              Supports ~33k normalizations or ~1k variable detections per day
                              Sufficient for 100k+ users
@@ -80,7 +80,7 @@ def easy_recurring_detection(
         >>> # V2: LLM-enhanced detection (better accuracy, minimal cost)
         >>> detector = easy_recurring_detection(enable_llm=True)
         >>> patterns = detector.detect_patterns(transactions)
-        >>> # Merchant normalization: "NFLX*SUB" → "Netflix" (90-95% accuracy)
+        >>> # Merchant normalization: "NFLX*SUB" -> "Netflix" (90-95% accuracy)
         >>> # Variable detection: Utility bills with seasonal variance (85-88% accuracy)
         >>> # Cost: ~$0.003/user/year with caching
 

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/recurring/insights.py

@@ -7,7 +7,7 @@ Provides on-demand insights for users:
 - Cost-saving recommendations (bundle deals, unused subscriptions)
 
 Uses ai-infra LLM with few-shot prompting.
-Caches results for 24 hours (80% hit rate expected) → <1ms latency.
+Caches results for 24 hours (80% hit rate expected) -> <1ms latency.
 Triggered via GET /recurring/insights API endpoint (not automatic).
 """
 
@@ -105,23 +105,23 @@ Guidelines:
 
 Examples:
 1. Subscriptions: Netflix $15.99, Hulu $12.99, Disney+ $10.99, Spotify $9.99, Amazon Prime $14.99
-   → "You have 5 subscriptions totaling $64.95/month. Consider the Disney+ bundle
+   -> "You have 5 subscriptions totaling $64.95/month. Consider the Disney+ bundle
       (Disney+, Hulu, ESPN+ for $19.99) to save $29.98/month. Also, Amazon Prime
       includes Prime Video - you may be able to cancel Netflix or Hulu."
-   → total_monthly_cost: 64.95
-   → potential_savings: 30.00
+   -> total_monthly_cost: 64.95
+   -> potential_savings: 30.00
 
 2. Subscriptions: Spotify $9.99, Apple Music $10.99
-   → "You're paying for both Spotify and Apple Music ($20.98/month). Cancel one
+   -> "You're paying for both Spotify and Apple Music ($20.98/month). Cancel one
       to save $10.99/month."
-   → total_monthly_cost: 20.98
-   → potential_savings: 10.99
+   -> total_monthly_cost: 20.98
+   -> potential_savings: 10.99
 
 3. Subscriptions: LA Fitness $40, Planet Fitness $10
-   → "You have 2 gym memberships totaling $50/month. Consider consolidating to
+   -> "You have 2 gym memberships totaling $50/month. Consider consolidating to
       just Planet Fitness to save $40/month."
-   → total_monthly_cost: 50.00
-   → potential_savings: 40.00
+   -> total_monthly_cost: 50.00
+   -> potential_savings: 40.00
 
 Output format (JSON):
 {
@@ -152,8 +152,8 @@ class SubscriptionInsightsGenerator:
     LLM-based subscription insights generator with caching.
 
     Layer 5 of 4-layer hybrid architecture (on-demand, optional):
-    1. Check cache first (80% hit rate, 24h TTL) → <1ms
-    2. Call LLM if cache miss → 300-500ms
+    1. Check cache first (80% hit rate, 24h TTL) -> <1ms
+    2. Call LLM if cache miss -> 300-500ms
     3. Cache result for 24 hours
     4. Return SubscriptionInsights
 
@@ -236,9 +236,9 @@ class SubscriptionInsightsGenerator:
         Generate subscription insights with natural language recommendations.
 
         Flow:
-        1. Check cache (80% hit rate, key: insights:{user_id}) → <1ms
+        1. Check cache (80% hit rate, key: insights:{user_id}) -> <1ms
         2. Check budget (daily/monthly caps)
-        3. Call LLM if cache miss → 300-500ms
+        3. Call LLM if cache miss -> 300-500ms
         4. Cache result (24h TTL)
         5. Return SubscriptionInsights
 

{fin_infra-0.2.1 → fin_infra-0.2.2}/src/fin_infra/recurring/normalizer.py

@@ -26,12 +26,12 @@ def normalize_merchant(raw_name: str) -> str:
     Normalize merchant name for grouping.
 
     Pipeline:
-    1. Lowercase: "NETFLIX.COM" → "netflix.com"
-    2. Remove domain suffixes: "netflix.com" → "netflix"
-    3. Remove special chars: "netflix*subscription" → "netflix subscription"
-    4. Remove store/transaction numbers: "starbucks #12345" → "starbucks"
-    5. Remove legal entities: "netflix inc" → "netflix"
-    6. Strip whitespace: "  netflix  " → "netflix"
+    1. Lowercase: "NETFLIX.COM" -> "netflix.com"
+    2. Remove domain suffixes: "netflix.com" -> "netflix"
+    3. Remove special chars: "netflix*subscription" -> "netflix subscription"
+    4. Remove store/transaction numbers: "starbucks #12345" -> "starbucks"
+    5. Remove legal entities: "netflix inc" -> "netflix"
+    6. Strip whitespace: "  netflix  " -> "netflix"
 
     Args:
         raw_name: Original merchant name