infrastructure-ai

_No description in Cargo.toml._

What it is for

Library crates for AI provider integration.

Provides a pluggable interface for AI backends with support for multiple providers, request/response types, and a stub provider for testing.

# Features

• AiProvider trait: Pluggable backend interface
• CompletionRequest/Response: Standard request/response types
• ProviderRegistry: Multi-provider management
• StubAiProvider: Testing and development provider

# Cargo Features

• http - Enable HTTP-based providers (OpenRouter, Anthropic)
• postgres - Enable encrypted credential storage
• full - Enable all features

# Example

use infrastructure_ai::{AiProvider, CompletionRequest, StubAiProvider};

# async fn example() -> infrastructure_ai::Result<()> {
// Use the stub provider for testing
let provider = StubAiProvider::new(true);

// Create a completion request
let request = CompletionRequest::new(
"You are a helpful assistant.",
"What is 2 + 2?"
)
.with_max_tokens(100)
.with_temperature(0.5);

// Get a response
let response = provider.complete(request).await?;
println!("Response: {}", response.content);

// Check token usage
if let Some(usage) = response.usage {
println!("Tokens used: {}", usage.total_tokens);
}
# Ok(())
# }

# Provider Registry

use infrastructure_ai::{ProviderRegistry, StubAiProvider};
use std::sync::Arc;

let mut registry = ProviderRegistry::new();
registry.register(Arc::new(StubAiProvider::new(true)));

// Get an available provider
if let Some(provider) = registry.available_provider() {
println!("Using provider: {}", provider.name());
}

# Per-Tenant AI (with full feature)

use infrastructure_ai::{AiProviderFactory, TenantAiSettings, AiProviderType};

// Create factory with database pool
let factory = AiProviderFactory::new(pool);

// Load tenant settings (from preferences or database)
let settings = TenantAiSettings::new(AiProviderType::OpenRouter)
.with_model("anthropic/claude-3.5-sonnet")
.with_enabled(true);

// Get provider for tenant (loads encrypted API key from database)
let provider = factory.for_tenant(tenant_id, &settings).await;

// Use provider
let response = provider.complete(request).await?;

Capabilities

answer (other)

Reading a model's answer back as JSON.

connection_setup (other)

Connection-setup helper for ai_credentials encryption (Sprint 0.12).

How to use it

From this crate's own rustdoc:

use infrastructure_ai::{AiProvider, CompletionRequest, StubAiProvider};

# async fn example() -> infrastructure_ai::Result<()> {
// Use the stub provider for testing
let provider = StubAiProvider::new(true);

// Create a completion request
let request = CompletionRequest::new(
    "You are a helpful assistant.",
    "What is 2 + 2?"
)
.with_max_tokens(100)
.with_temperature(0.5);

// Get a response
let response = provider.complete(request).await?;
println!("Response: {}", response.content);

// Check token usage
if let Some(usage) = response.usage {
    println!("Tokens used: {}", usage.total_tokens);
}
# Ok(())
# }

From this crate's own rustdoc:

use infrastructure_ai::{ProviderRegistry, StubAiProvider};
use std::sync::Arc;

let mut registry = ProviderRegistry::new();
registry.register(Arc::new(StubAiProvider::new(true)));

// Get an available provider
if let Some(provider) = registry.available_provider() {
    println!("Using provider: {}", provider.name());
}

Module structure

infrastructure_ai

• answer
• connection_setup

flowchart TD
  n_infrastructure_ai["infrastructure_ai"]
  n_infrastructure_ai --> n_answer["answer"]
  n_infrastructure_ai --> n_connection_setup["connection_setup"]

Public surface

`answer`

`connection_setup`

Re-exports. Exported here, defined elsewhere.

Boundary

Reaches into foundation.

Shares tier infrastructure with 82 other crates: infrastructure-acquire, infrastructure-adapters-google-calendar, infrastructure-adapters-google-gmail, infrastructure-adapters-google-places, infrastructure-adapters-google-trends, infrastructure-adapters-shodan, infrastructure-adapters-yelp, infrastructure-agent, … (82 total).

_What this crate deliberately does NOT own is a judgment. No committed registry records one for it, so none is stated here._

Where it sits

Tier flow. Which tiers this crate's own edges cross.

flowchart LR
  n_infrastructure["infrastructure"] --> n_foundation["foundation"]

Dependencies

Runtime, in this workspace.

Runtime, from outside the workspace.

Development, from outside the workspace.

Build. None.

Depended on by. 17 workspace crates.

Signal flow — what reaches this crate, and what it reaches.

flowchart LR
  n_application_ai["application-ai"] -->|uses| SELF
  n_application_classify["application-classify"] -->|uses| SELF
  n_application_cms["application-cms"] -->|uses| SELF
  n_application_conversation["application-conversation"] -->|uses| SELF
  n_application_messaging["application-messaging"] -->|uses| SELF
  n_domain_ai_report["domain-ai-report"] -->|uses| SELF
  n_infrastructure_agent["infrastructure-agent"] -->|uses| SELF
  n_infrastructure_ocr_ai["infrastructure-ocr-ai"] -->|uses| SELF
  n_infrastructure_ocr_provider["infrastructure-ocr-provider"] -->|uses| SELF
  n_infrastructure_translate["infrastructure-translate"] -->|uses| SELF
  n_platform_privacy_scan_api["platform-privacy-scan-api"] -->|uses| SELF
  n_tools_doc_ask["tools-doc-ask"] -->|uses| SELF
  n_tools_imp["tools-imp"] -->|uses| SELF
  n_tools_knowitall["tools-knowitall"] -->|uses| SELF
  n_tools_local_ask["tools-local-ask"] -->|uses| SELF
  n_tools_packet_compile["tools-packet-compile"] -->|uses| SELF
  n_tools_prompt_shaper["tools-prompt-shaper"] -->|uses| SELF
  SELF["infrastructure-ai"]
  SELF -->|runtime| n_foundation_bounded_io["foundation-bounded-io"]
  SELF -->|runtime| n_infrastructure_fetcher["infrastructure-fetcher"]
  classDef self fill:#1f883d,stroke:#1f883d,color:#fff;
  class SELF self;

Feature flags

flowchart LR
  n_cli["cli"] --> n_tokio["tokio"]
  n_default["default"]
  n_full["full"] --> n_http["http"]
  n_full["full"] --> n_postgres["postgres"]
  n_full["full"] --> n_cli["cli"]
  n_http["http"] --> n_reqwest["reqwest"]
  n_http["http"] --> n_dep_infrastructure_fetcher["dep:infrastructure-fetcher"]
  n_postgres["postgres"] --> n_dep_sqlx["dep:sqlx"]
  n_postgres["postgres"] --> n_dep_chrono["dep:chrono"]
  n_reqwest["reqwest"] --> n_dep_reqwest["dep:reqwest"]
  n_tokio["tokio"] --> n_dep_tokio["dep:tokio"]

Targets

Error model

No public error type was detected: no public item declares a type named *Error, and no public signature returns one.

Operational characteristics

No unsafe block, unsafe fn, unsafe impl or unsafe trait was found by the parser anywhere in this crate's source.

Configuration

Related capabilities

17 workspace crates depend on this one: application-ai, application-classify, application-cms, application-conversation, application-messaging, domain-ai-report, infrastructure-agent, infrastructure-ocr-ai, infrastructure-ocr-provider, infrastructure-translate, platform-privacy-scan-api, tools-doc-ask, … (17 total).

Verification

Evidence by module. How often each public module is named by something executable.

What the tests establish, by name:

• after_connect_hook_lets_inserts_succeed — tests/connection_setup_smoke.rs
• without_hook_inserts_are_rejected — tests/connection_setup_smoke.rs
• a_bare_object_is_read_as_it_stands — src/answer.rs
• a_fence_around_the_object_is_not_part_of_the_object — src/answer.rs
• a_missing_field_is_still_missing_after_unwrapping — src/answer.rs
• a_result_that_is_not_a_string_is_left_alone — src/answer.rs
• an_answer_with_no_object_says_so_rather_than_guessing — src/answer.rs
• json_as_reads_the_unwrapped_object_into_the_callers_type — src/answer.rs
• prose_on_either_side_of_the_object_is_ignored — src/answer.rs
• something_between_braces_that_is_not_json_is_reported_as_not_json — src/answer.rs
• the_answer_inside_a_worker_envelope_is_the_answer — src/answer.rs
• debug_does_not_leak_api_key — src/anthropic.rs
• test_available_models — src/anthropic.rs
• test_empty_key_not_available — src/anthropic.rs
• test_image_content_serializes_as_blocks — src/anthropic.rs
• test_image_only_content_serializes_single_block — src/anthropic.rs
• test_provider_creation — src/anthropic.rs
• test_text_only_content_serializes_as_string — src/anthropic.rs
• assistant_tool_calls_has_empty_content — src/chat.rs
• capabilities_builder — src/chat.rs
• request_builder_sets_fields — src/chat.rs
• response_accessors — src/chat.rs
• tool_calls_skip_serializing_when_empty — src/chat.rs
• tool_result_roundtrips_role_and_name — src/chat.rs
• claude_cli_real_translation — src/claude_cli.rs
• claude_cli_real_vision_ocr — src/claude_cli.rs
• combine_prompt_does_not_escape_or_alter_metacharacters — src/claude_cli.rs
• combine_prompt_joins_system_and_user — src/claude_cli.rs
• malicious_input_never_reaches_a_shell — src/claude_cli.rs
• truncate_bounded_respects_cap_and_codepoints — src/claude_cli.rs
• _… 42 more_

Documentation coverage

pie showData
    title Public items with rustdoc
    "Documented" : 4
    "No rustdoc detected" : 0

Metrics

pie showData
    title Public API by kind
    "function" : 4

pie showData
    title Rust source composition
    "Code" : 3089
    "Blank or comment" : 1306

Generation

Rendered by tools-corpus corpus readme from repository evidence alone, renderer schema 2, lexicon current. No model, network service or database was consulted. Regenerate with tools-corpus corpus readme --write; verify with --check.