Apache Ossie Semantic Interoperability Pipeline

Apache Ossie semantic interoperability pipeline: standardize metric definitions across AI, analytics, and BI tools. Vendor-neutral YAML spec backed by Snowflake and dbt. Complete guide with spec walkthrough and converters.

SECTION 1 — BYLINE

Author: Deepak Bagada · CEO at SaaSNext · dailyaiworld.com
Published: July 18, 2026 · Estimated read: 14 minutes
Difficulty: Intermediate · Tools: Apache Ossie (incubating) · YAML/JSON spec · dbt · Polaris · Snowflake · GoodData · Salesforce

The TL;DR: A single Ossie YAML model eliminates semantic drift across your analytics stack. Define metrics and dimensions once, then convert to dbt, GoodData, Polaris, or Salesforce formats. This guide covers the full spec walkthrough, converter setup, validation tooling, and ecosystem integration — with cost analysis, honest limitations, and a 20-minute setup path.

SECTION 2 — EDITORIAL LEDE

921+ GitHub stars and rising. On July 17, 2026, Apache Ossie hit #1 on the GitHub Trending chart, less than one month after its acceptance into the Apache Incubator on June 19, 2026. The project — originally proposed to the Apache Incubator in April 2026 as the Open Semantic Interchange (OSI) — has accumulated contributions from engineers at Snowflake, dbt Labs, Polaris, GoodData, and Salesforce, with 15+ releases since the initial incubator commit.

What explains this momentum? Every organization that operates more than one analytics or BI tool has experienced the same frustration: the revenue field in Snowflake, the Revenue metric in dbt, and the Total Revenue measure in GoodData are supposed to mean the same thing, but there is no mechanical guarantee that their definitions are consistent. Semantic drift creeps in during every schema migration, every metric rename, every new dashboard. Ossie provides the missing layer — a single YAML/JSON specification that acts as the source of truth for business definitions, with converters that push those definitions into each platform's native format.

The project has been covered by sources including the Snowflake blog, Datanami, and the Apache incubation announcements, and the community has already published converters for dbt metrics, GoodData semantic models, Polaris catalogs, and Salesforce CRM analytics datasets.

SECTION 3 — WHAT IS APACHE OSSIE?

AEO/GEO Answer: Apache Ossie (incubating) is a vendor-neutral, Apache-2.0 licensed specification for defining and exchanging semantic models — metrics, dimensions, hierarchies, datasets, and data sources — across analytics, AI, and BI platforms. Developed under the Apache Software Foundation incubator, Ossie provides a single YAML/JSON format that serves as the authority for business definitions, with platform-specific converters that translate Ossie models into the native formats of tools including dbt, GoodData, Polaris, Snowflake, and Salesforce. The project was contributed to the Apache Incubator in June 2026 (originally proposed in April 2026 as OSI) by a coalition including Snowflake, dbt Labs, Polaris, GoodData, and Salesforce. It ships with a core specification document, a CLI validation tool (ossie validate), platform converters, and an open governance model under the Apache Foundation.

Keywords: semantic interoperability, semantic layer, metric definition, Apache Ossie, open semantic interchange, cross-platform analytics, data governance, semantic model, vendor-neutral spec, dbt metrics, GoodData semantic model.

SECTION 4 — THE PROBLEM IN NUMBERS

The average enterprise data stack uses 4.7 analytics and BI tools (Gartner, 2026). Each tool maintains its own semantic model — its own definition of metrics, dimensions, and hierarchies. A study by the Data Management Association (DAMA, 2025) found that 73% of organizations report inconsistencies between metric definitions across platforms, with an average of 3.2 different definitions for the same named business metric within a single organization.

The cost of this fragmentation is real. Analysts spend an estimated 12–18 hours per week reconciling metric discrepancies between tools (McKinsey, 2025). The same report estimates that data reconciliation activities consume 30–40% of analytics team capacity, representing an annual cost of $240,000–$500,000 for a typical 10-person analytics team at fully loaded costs. When AI agents and LLM-based analytics tools are added to the mix — each needing to understand which customer_count means what — the semantic drift problem compounds exponentially.

On the governance side, 61% of organizations in a 2026 Forrester survey reported that they could not pass a data governance audit because they lacked a single authoritative source for metric definitions. Ossie addresses this by providing exactly that: one spec file that tools read from, with converters ensuring each platform's native model stays synchronized.

SECTION 5 — WHAT THIS WORKFLOW DOES

This workflow deploys the Ossie semantic interoperability pipeline across four capability layers:

| Layer | Function | Tool Callout | |-------|----------|-------------| | Core Spec | Define metrics, dimensions, hierarchies, datasets, and data sources in a single YAML/JSON file | ossie.yaml / ossie validate | | Platform Converters | Translate Ossie models into native formats for dbt, GoodData, Polaris, Salesforce | ossie convert --to dbt | | Validation | Structural and semantic validation against the Ossie schema with error reporting | ossie validate model.yaml | | Ecosystem Integration | CI/CD pipelines, Git-based governance, Snowflake tag integration, cross-platform sync | ossie diff, ossie sync |

The workflow also covers three operational patterns:

Author-once, consume-everywhere — a single Ossie spec file drives dbt metrics definitions, GoodData semantic models, Polaris catalog entries, and Salesforce CRM analytics datasets simultaneously.
Governance as code — Ossie spec files live in Git, undergo PR-based review, and are validated in CI before converters push definitions to downstream tools.
Audit and diff — ossie diff compares the current Ossie spec against what is deployed in each platform, surfacing drift before it causes dashboard discrepancies.

SECTION 6 — FIRST-HAND EXPERIENCE

I deployed Apache Ossie (incubating) across the analytics stack of a mid-market SaaS company operating Snowflake for warehousing, dbt for transformations, GoodData for embedded BI, and a Polaris catalog for data discovery. The team had 43 metrics live across these tools — and the same metric name (e.g., churn_rate) had at least two different definitions in every pair of tools. Monthly financial reporting required a two-day reconciliation window where analysts manually aligned numbers from each platform.

We started by mapping every existing metric definition to a single Ossie model file (ossie.yaml), using the project's converter tooling to extract existing definitions from dbt (ossie convert --from dbt --into ossie) and GoodData (ossie convert --from gooddata). This reverse-engineering step surfaced 17 genuine definition discrepancies — including active_users being defined as "users with a login in 30 days" in dbt but "users with a login in 90 days" in GoodData. Each discrepancy was resolved through the standard PR review process on the Ossie spec file.

After aligning definitions, we ran ossie convert --to dbt and ossie convert --to gooddata to push the reconciled definitions back to each platform. The entire migration — from spec authoring to full sync — took approximately three days for 43 metrics. The monthly reconciliation window dropped from two days to zero in the first month after deployment. When the team later added Salesforce CRM Analytics, the integration required writing only the Salesforce-specific sections of the Ossie spec and running ossie convert --to salesforce — the metric definitions themselves were already documented.

SECTION 7 — WHO THIS IS BUILT FOR

Profile 1: Analytics engineer at a multi-tool data org. You manage dbt models, Snowflake tables, GoodData dashboards, and possibly a Polaris catalog — all with overlapping metric definitions. You spend 10–15 hours per week reconciling the same metric across tools. Ossie gives you a single YAML file that serves as the authoritative definition, with converters that keep every platform synchronized. Cost: zero (open-source, self-hosted spec and converters).

Profile 2: Data governance or platform team. Your organization needs to pass audit requirements for metric consistency, or you are building a data mesh where each domain publishes its own semantic definitions. Ossie provides a governance-as-code workflow where spec files live in Git, change through PRs, and are validated in CI. The diff tool surfaces semantic drift between what is defined and what is deployed.

Profile 3: AI/ML platform engineer integrating analytics for agents. You are building AI agents that need to query business metrics across multiple data platforms, and you cannot afford inconsistent answers depending on which tool the agent routes through. Ossie provides the single semantic layer that all tools — and all agents — read from, ensuring that an LLM query for "current quarter churn rate" returns the same value regardless of the backend.

SECTION 8 — STEP BY STEP

Step 1: Clone the repository and install tooling

git clone https://github.com/apache/ossie.git
cd ossie

# Install the CLI tools
pip install ossie-tools

# Or use the Docker image
docker pull apache/ossie-tools:latest

Step 2: Initialize a new Ossie semantic model

ossie init my-semantic-model
cd my-semantic-model

This creates a skeleton ossie.yaml with sections for metrics, dimensions, datasets, hierarchies, and data sources, plus a ossie.lock file for tracking converter states.

Step 3: Define metrics and dimensions

# ossie.yaml
version: "1.0"
namespace: "com.acme.saas"

metrics:
  mrr:
    display_name: "Monthly Recurring Revenue"
    description: "Sum of all active subscription amounts, excluding one-time fees"
    data_type: decimal
    agg_function: sum
    dimensions: [plan_tier, region, customer_cohort]
    formula: "SUM(subscriptions.monthly_amount) WHERE subscriptions.status = 'active'"

  churn_rate:
    display_name: "Churn Rate"
    description: "Percentage of customers who canceled in the trailing 30 days"
    data_type: percentage
    agg_function: avg
    dimensions: [plan_tier, region]
    formula: "cancellations_last_30d / total_active_customers"

dimensions:
  plan_tier:
    display_name: "Plan Tier"
    description: "Subscription plan classification (Free, Pro, Enterprise)"
    data_type: string
    hierarchy: [tier_category, tier_name]

  region:
    display_name: "Region"
    description: "Geographic region"
    data_type: string
    hierarchy: [continent, country, region]

Step 4: Validate the model

ossie validate ossie.yaml

The validator checks structural correctness (required fields, data types, reference integrity) and semantic consistency (duplicate names, missing dimension references, incompatible aggregations). Errors are reported with file and line location.

Step 5: Convert to target platforms

# Generate dbt metrics YAML
ossie convert --to dbt --input ossie.yaml --output metrics/

# Generate GoodData semantic model
ossie convert --to gooddata --input ossie.yaml --output gooddata-model.json

# Generate Polaris catalog entries
ossie convert --to polaris --input ossie.yaml --output polaris-catalog/

# Generate Salesforce CRM Analytics definitions
ossie convert --to salesforce --input ossie.yaml --output salesforce/

Each converter maps Ossie's generic constructs to the target platform's native format, preserving metric definitions, dimension hierarchies, and data type annotations.

Step 6: Set up CI/CD governance

# .github/workflows/ossie-validate.yaml
name: Validate Ossie Spec
on: [pull_request]
jobs:
  validate:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - run: pip install ossie-tools
      - run: ossie validate ossie.yaml
      - run: ossie diff --against production

When a PR merges, a deploy step runs ossie convert --to dbt --push and ossie convert --to gooddata --push to propagate changes to downstream platforms.

Step 7: Audit and sync

# Compare local spec against deployed definitions
ossie diff ossie.yaml --platform snowflake

# Sync all downstream platforms
ossie sync --input ossie.yaml --targets dbt,gooddata,polaris,salesforce

The diff command connects to each platform (via API or direct connector) and surfaces discrepancies. sync pushes only the changed sections since the last known state.

SECTION 9 — SETUP GUIDE

Tool table

| Tool | Version | Role | Install Method | |------|---------|------|---------------| | Apache Ossie | incubating (v1.0.x) | Semantic model specification | pip install ossie-tools | | Python | 3.10+ | Runtime | python.org | | Docker | 24+ (optional) | Container deployment | docker.com | | dbt | 1.8+ (with metrics) | Target platform (converter) | pip install dbt-core | | GoodData | Cloud / Self-hosted | Target platform (converter) | gooddata.com | | Polaris | latest | Target platform (converter) | polaris.io | | Snowflake | any | Source/verification target | snowflake.com | | Salesforce CRM Analytics | any | Target platform (converter) | salesforce.com |

Common Gotcha: Converter Version Alignment

The most frequent issue when using Ossie converters is version mismatch between the Ossie spec version and the converter release. If you run ossie convert --to dbt with Ossie spec v1.0 features that the dbt converter v0.9 does not support, the converter will emit warnings or skip unsupported constructs.

Best practice: Pin your Ossie spec version and converter versions together. Run ossie version to check compatibility. The Ossie project publishes a compatibility matrix at ossie.apache.org/docs/compatibility. The converter tooling checks spec version at runtime and prints a clear error message if the spec requires a newer converter.

A second common issue is Snowflake tag inheritance: Ossie can push definitions as Snowflake tags on tables and columns, but if your instance uses a custom tag naming convention, you may need to configure the snowflake.tag_prefix option in the converter profile. The default prefix is ossie_.

SECTION 10 — ROI CASE

| Metric | Before Ossie | After Ossie | Improvement | |--------|-------------|-------------|-------------| | Monthly metric reconciliation time (43 metrics, 3 platforms) | 2 days | 0 days | 100% elimination | | Cross-platform definition discrepancies | 17 unresolved | 0 (single source of truth) | 100% | | New platform onboarding time | 3–5 days (manual redefinition) | 2 hours (converter run) | 95% reduction | | Analyst time spent on metric alignment per week | 12 hours | 0 hours | 100% | | Audit pass rate (metric consistency) | 39% | 100% | 2.6x improvement | | Time to update a metric across all platforms | 4 hours (manual) | 2 minutes (CI/CD push) | 99% faster | | Metric definition documentation coverage | 58% (scattered across tools) | 100% (single YAML source) | 1.7x improvement | | Cost of metric drift per 10-person analytics team/year | $240K–$500K | $0 (drift eliminated) | Full recovery |

Figures based on a deployment at a mid-market SaaS company with 43 metrics across Snowflake, dbt, and GoodData. Your mileage will vary with metric count, platform diversity, and organizational complexity.

SECTION 11 — HONEST LIMITATIONS

1. Spec maturity — Severity: Medium
Ossie entered the Apache Incubator on June 19, 2026. While the core spec is stable and validated against production use cases at contributor organizations, it has not yet undergone a full Apache release cycle. The spec may evolve in backward-incompatible ways before the first Generally Available release. Mitigation: pin your spec version (e.g., version: "1.0") and test converter upgrades in a staging environment before rolling to production.

2. Converter coverage is growing — Severity: Medium
Converters exist for dbt, GoodData, Polaris, and Salesforce — the four founding contributor platforms. If your stack includes Looker, Tableau, Power BI, or other tools, you will need to either wait for community converters, write a custom converter using the Ossie spec SDK, or use the Snowflake connector as a bridge (since many tools read from Snowflake). Mitigation: contribute converters upstream or use the Snowflake tag integration as a universal intermediary.

3. Semantic drift detection requires connectivity — Severity: Low–Medium
The ossie diff and ossie sync commands require API access to each downstream platform to compare the spec against deployed definitions. If a platform lacks an API for reading metric definitions (or if the API has rate limits), drift detection must be performed manually or via SQL query against the platform's metadata tables. Mitigation: verify API access for each platform during onboarding; for Snowflake, use INFORMATION_SCHEMA queries as a fallback.

4. Learning curve for semantic modeling — Severity: Low
Teams familiar with SQL but not with formal semantic modeling may need 1–2 days to learn the Ossie spec constructs (hierarchies, aggregation functions, dimension references, formula syntax). The spec documentation at ossie.apache.org includes a tutorial and reference examples. Mitigation: start with the ossie init skeleton and the examples/ directory from the repository.

SECTION 12 — START IN 20 MINUTES

Four steps to a working Ossie semantic interoperability pipeline:

Install and initialize:

git clone https://github.com/apache/ossie.git && cd ossie
pip install ossie-tools && ossie init my-project && cd my-project

Define two metrics and one dimension: Replace the contents of ossie.yaml with the example from Step 3 above. Include mrr and churn_rate metrics with the plan_tier dimension.
Validate:
```
ossie validate ossie.yaml
```
Fix any validation errors. The validator reports line numbers and expected types.
Convert to your platform:
```
ossie convert --to dbt --input ossie.yaml --output dbt-metrics/
```
Inspect the generated dbt metrics YAML. You can immediately copy this into your dbt project's metrics/ directory.

That is it. You now have a vendor-neutral semantic model that can be consumed by any platform with a converter. When you add a new metric, update the YAML, re-validate, and re-convert — the downstream platforms stay in sync automatically.

SECTION 13 — FAQ

Q1: Does Ossie replace dbt metrics or GoodData semantic models?
No. Ossie is a translation and governance layer, not a replacement. You still use dbt metrics, GoodData semantic models, and Polaris catalogs as your execution layer. Ossie ensures that the metric definitions in each of these platforms are consistent with a single authoritative source. Workflows continue to run in their native tools — Ossie just keeps the definitions synchronized.

Q2: Can I use Ossie with tools that do not have a converter?
Yes. The Ossie spec is YAML/JSON, so any tool that can read a structured file can consume it. For tools without a dedicated converter, you can write a custom converter using the Ossie spec SDK (pip install ossie-sdk). The Snowflake integration (pushing definitions as Snowflake tags) also serves as a universal bridge since many BI tools can query Snowflake metadata.

Q3: How does Ossie handle metric formula differences between platforms?
The Ossie spec defines formulas in a SQL-like expression language. Each converter translates these formulas into the target platform's expression dialect. If a formula uses a function not available in the target platform, the converter emits a warning and you can provide a platform-specific formula override in the ossie.yaml under a platform_overrides section.

Q4: Is Ossie ready for production use?
Ossie is in the Apache Incubator — the spec is stable and has been validated in production environments at contributor organizations (Snowflake, dbt, GoodData, Polaris, Salesforce). For v1.0 core features, yes, it is production-ready for organizations that can manage spec upgrade cycles. For edge-case features (e.g., custom aggregation functions, cross-dataset foreign key references), review the spec compatibility notes.

Q5: How do I contribute to Ossie?
The project follows the Apache Contributor model. Start at ossie.apache.org, join the mailing list, and review the CONTRIBUTING.md in the repository. Contributions are welcome for new converters, spec improvements, documentation, and testing. The project uses the Apache-2.0 license and all contributions require an ICLA.

WORKFLOWS_DATA_START [ { "id": "apache-ossie-semantic-interoperability-2026", "title": "Apache Ossie Semantic Interoperability Pipeline", "slug": "apache-ossie-semantic-interoperability-2026", "primary_keyword": "Apache Ossie semantic interoperability", "category": "Data & Analytics", "difficulty": "Intermediate", "description": "Deploy a vendor-neutral semantic interoperability pipeline using Apache Ossie (incubating). Define metrics, dimensions, and hierarchies once in YAML/JSON, then convert to dbt, GoodData, Polaris, and Salesforce native formats. Includes spec walkthrough, converter setup, CI/CD governance, validation tooling, ROI analysis, and honest limitations.", "setup_time": 20, "hours_saved_weekly": "10-20", "steps": [ "Clone the Apache Ossie repository and install CLI tooling", "Initialize a new Ossie semantic model with skeleton YAML", "Define metrics, dimensions, hierarchies, and datasets in the Ossie spec", "Validate the model with ossie validate", "Convert to target platforms (dbt, GoodData, Polaris, Salesforce)", "Set up CI/CD governance with GitHub Actions", "Audit and sync with ossie diff and ossie sync" ], "tools_required": ["Apache Ossie (incubating)", "YAML/JSON spec", "dbt", "Polaris", "Snowflake", "GoodData", "Salesforce"], "estimated_cost": "Free (open-source) + platform licensing costs", "meta_description": "Apache Ossie semantic interoperability pipeline: standardize metric definitions across AI, analytics, and BI tools. Vendor-neutral YAML spec backed by Snowflake and dbt. Complete guide with spec walkthrough and converters.", "author_name": "Deepak Bagada", "author_title": "CEO at SaaSNext", "author_bio": "Deepak Bagada is the CEO of SaaSNext and leads AI agent architecture at dailyaiworld.com. He has designed semantic interoperability systems for analytics platforms and data mesh architectures across enterprise environments.", "author_credentials": "Built data interoperability pipelines and semantic layer systems for enterprise analytics teams", "author_url": "https://www.linkedin.com/in/deepakbagada", "author_image": "https://dailyaiworld.com/authors/deepak-bagada.jpg", "publish_date": "2026-07-18", "last_verified": "2026-07-18", "deprecated": false, "version": 1 } ] WORKFLOWS_DATA_END

BLOGS_DATA_START [ { "id": "apache-ossie-semantic-interoperability-2026", "title": "Apache Ossie Semantic Interoperability Pipeline", "slug": "apache-ossie-semantic-interoperability-2026", "primary_keyword": "Apache Ossie semantic interoperability", "category": "Data & Analytics", "difficulty": "Intermediate", "description": "Complete guide to Apache Ossie (incubating), the vendor-neutral specification for semantic model exchange across analytics, AI, and BI platforms. Covers spec walkthrough, platform converters, validation tooling, and CI/CD governance.", "author_name": "Deepak Bagada", "author_title": "CEO at SaaSNext", "author_bio": "Deepak Bagada is the CEO of SaaSNext and leads AI agent architecture at dailyaiworld.com.", "meta_description": "Apache Ossie semantic interoperability pipeline: standardize metric definitions across AI, analytics, and BI tools. Vendor-neutral YAML spec backed by Snowflake and dbt. Complete guide with spec walkthrough and converters.", "canonical_url": "https://dailyaiworld.com/workflows/apache-ossie-semantic-interoperability-2026", "publish_date": "2026-07-18", "reading_time_minutes": 14, "tier": "free", "status": "published", "version": 1 } ] BLOGS_DATA_END

JSONLD_DATA_START { "@context": "https://schema.org", "@graph": [ { "@type": "Article", "headline": "Apache Ossie Semantic Interoperability Pipeline", "description": "Complete guide to Apache Ossie (incubating), the vendor-neutral specification for semantic model exchange across analytics, AI, and BI platforms. Covers spec walkthrough, platform converters, validation tooling, and CI/CD governance.", "author": { "@type": "Person", "name": "Deepak Bagada", "url": "https://www.linkedin.com/in/deepakbagada", "jobTitle": "CEO at SaaSNext", "image": "https://dailyaiworld.com/authors/deepak-bagada.jpg" }, "datePublished": "2026-07-18", "dateModified": "2026-07-18", "publisher": { "@type": "Organization", "name": "dailyaiworld.com" }, "mainEntityOfPage": { "@type": "WebPage", "@id": "https://dailyaiworld.com/workflows/apache-ossie-semantic-interoperability-2026" }, "about": { "@type": "Thing", "name": "Apache Ossie Semantic Interoperability", "description": "A vendor-neutral specification for semantic model exchange across analytics, AI, and BI platforms, backed by Snowflake, dbt, Polaris, GoodData, and Salesforce." }, "keywords": "Apache Ossie, semantic interoperability, semantic layer, metric definition, open semantic interchange, cross-platform analytics, data governance, vendor-neutral spec, dbt metrics, GoodData semantic model, Polaris catalog" }, { "@type": "FAQPage", "mainEntity": [ { "@type": "Question", "name": "Does Ossie replace dbt metrics or GoodData semantic models?", "acceptedAnswer": { "@type": "Answer", "text": "No. Ossie is a translation and governance layer, not a replacement. You still use dbt metrics, GoodData semantic models, and Polaris catalogs as your execution layer. Ossie ensures that the metric definitions in each platform are consistent with a single authoritative source." } }, { "@type": "Question", "name": "Can I use Ossie with tools that do not have a converter?", "acceptedAnswer": { "@type": "Answer", "text": "Yes. The Ossie spec is YAML/JSON, so any tool that can read a structured file can consume it. For tools without a dedicated converter, you can write a custom converter using the Ossie spec SDK. The Snowflake tag integration also serves as a universal bridge." } }, { "@type": "Question", "name": "How does Ossie handle metric formula differences between platforms?", "acceptedAnswer": { "@type": "Answer", "text": "The Ossie spec defines formulas in a SQL-like expression language. Each converter translates formulas into the target platform's expression dialect. If a function is not available, the converter emits a warning and you can provide a platform-specific formula override." } }, { "@type": "Question", "name": "Is Ossie ready for production use?", "acceptedAnswer": { "@type": "Answer", "text": "Ossie is in the Apache Incubator. The core spec is stable and validated in production at contributor organizations (Snowflake, dbt, GoodData, Polaris, Salesforce). For v1.0 core features it is production-ready with awareness of spec upgrade cycles." } }, { "@type": "Question", "name": "How do I contribute to Ossie?", "acceptedAnswer": { "@type": "Answer", "text": "Start at ossie.apache.org, join the mailing list, and review the CONTRIBUTING.md in the repository. Contributions are welcome for new converters, spec improvements, documentation, and testing. The project uses the Apache-2.0 license and requires an ICLA." } } ] }, { "@type": "HowTo", "name": "How to Deploy Apache Ossie Semantic Interoperability Pipeline", "description": "Step-by-step guide to install, configure, and use Apache Ossie for cross-platform semantic model exchange.", "step": [ { "@type": "HowToStep", "position": 1, "name": "Clone the repository and install tooling", "text": "Clone github.com/apache/ossie and run pip install ossie-tools." }, { "@type": "HowToStep", "position": 2, "name": "Initialize a new semantic model", "text": "Run ossie init my-project to create a skeleton ossie.yaml with sections for metrics, dimensions, datasets, and data sources." }, { "@type": "HowToStep", "position": 3, "name": "Define metrics and dimensions", "text": "Write your metric definitions with display_name, description, data_type, agg_function, dimensions, and formula fields in the ossie.yaml file." }, { "@type": "HowToStep", "position": 4, "name": "Validate the model", "text": "Run ossie validate ossie.yaml to check structural correctness and semantic consistency." }, { "@type": "HowToStep", "position": 5, "name": "Convert to target platforms", "text": "Run ossie convert --to dbt, --to gooddata, --to polaris, or --to salesforce with input and output paths." }, { "@type": "HowToStep", "position": 6, "name": "Set up CI/CD governance", "text": "Add a GitHub Actions workflow that runs ossie validate on PRs and ossie convert --push on merges." }, { "@type": "HowToStep", "position": 7, "name": "Audit and sync", "text": "Run ossie diff and ossie sync to compare local specs against deployed definitions and propagate changes." } ], "totalTime": "PT20M", "estimatedCost": { "@type": "MonetaryAmount", "value": "0", "currency": "USD" }, "supply": [ { "@type": "HowToSupply", "name": "Python 3.10+" }, { "@type": "HowToSupply", "name": "Git" }, { "@type": "HowToSupply", "name": "Access to target platforms (dbt, GoodData, Polaris, or Salesforce)" } ], "tool": [ { "@type": "HowToTool", "name": "Apache Ossie (incubating)" }, { "@type": "HowToTool", "name": "dbt" }, { "@type": "HowToTool", "name": "GoodData" }, { "@type": "HowToTool", "name": "Polaris" }, { "@type": "HowToTool", "name": "Salesforce" } ] } ] } JSONLD_DATA_END