feat: add fact tables and issue-label bridge to the dimensional model

dbt/models/marts/_marts__models.yml

@@ -2,7 +2,7 @@ version: 2
 
 models:
   - name: dim_dates
-    description: Calendar dimension, one row per day (2020–2030), generated in SQL.
+    description: Calendar dimension, one row per day (2020-2030), generated in SQL.
     columns:
       - name: date_key
         description: Surrogate date key (YYYYMMDD integer).
@@ -43,3 +43,162 @@ models:
         data_tests: [not_null, unique]
       - name: label_name
         data_tests: [not_null, unique]
+
+  - name: fct_commits
+    description: One row per commit. Repository resolved via the SCD2 half-open date range.
+    columns:
+      - name: commit_key
+        description: Surrogate key, hash of repository + commit_sha.
+        data_tests: [not_null, unique]
+      - name: repository_key
+        data_tests:
+          - not_null
+          - relationships:
+              arguments:
+                to: ref('dim_repositories')
+                field: repository_key
+      - name: contributor_key
+        description: Nullable - commits whose author has no GitHub login stay unattributed.
+        data_tests:
+          - relationships:
+              arguments:
+                to: ref('dim_contributors')
+                field: contributor_key
+      - name: date_key
+        data_tests:
+          - not_null
+          - relationships:
+              arguments:
+                to: ref('dim_dates')
+                field: date_key
+      - name: lines_added
+        data_tests:
+          - dbt_utils.expression_is_true:
+              arguments:
+                expression: ">= 0"
+      - name: lines_deleted
+        data_tests:
+          - dbt_utils.expression_is_true:
+              arguments:
+                expression: ">= 0"
+
+  - name: fct_pull_requests
+    description: One row per pull request. Repository resolved via the SCD2 half-open range on the PR open date.
+    columns:
+      - name: pr_key
+        description: Surrogate key, hash of repository + pr_number.
+        data_tests: [not_null, unique]
+      - name: repository_key
+        data_tests:
+          - not_null
+          - relationships:
+              arguments:
+                to: ref('dim_repositories')
+                field: repository_key
+      - name: author_key
+        description: Nullable FK to dim_contributors.
+        data_tests:
+          - relationships:
+              arguments:
+                to: ref('dim_contributors')
+                field: contributor_key
+      - name: opened_date_key
+        data_tests:
+          - not_null
+          - relationships:
+              arguments:
+                to: ref('dim_dates')
+                field: date_key
+      - name: merged_date_key
+        description: Nullable - only set for merged PRs.
+        data_tests:
+          - relationships:
+              arguments:
+                to: ref('dim_dates')
+                field: date_key
+      - name: review_count
+        data_tests:
+          - dbt_utils.expression_is_true:
+              arguments:
+                expression: ">= 0"
+
+  - name: fct_issues
+    description: One row per issue. Repository resolved via the SCD2 half-open range.
+    columns:
+      - name: issue_key
+        description: Surrogate key, hash of repository + issue_number.
+        data_tests: [not_null, unique]
+      - name: repository_key
+        data_tests:
+          - not_null
+          - relationships:
+              arguments:
+                to: ref('dim_repositories')
+                field: repository_key
+      - name: author_key
+        description: Nullable FK to dim_contributors.
+        data_tests:
+          - relationships:
+              arguments:
+                to: ref('dim_contributors')
+                field: contributor_key
+      - name: opened_date_key
+        data_tests:
+          - not_null
+          - relationships:
+              arguments:
+                to: ref('dim_dates')
+                field: date_key
+      - name: closed_date_key
+        description: Nullable - only set for closed issues.
+        data_tests:
+          - relationships:
+              arguments:
+                to: ref('dim_dates')
+                field: date_key
+      - name: comment_count
+        data_tests:
+          - dbt_utils.expression_is_true:
+              arguments:
+                expression: ">= 0"
+
+  - name: bridge_issue_labels
+    description: Bridge for the issue<->label many-to-many. Grain - one row per issue/label.
+    data_tests:
+      - dbt_utils.unique_combination_of_columns:
+          arguments:
+            combination_of_columns: [issue_key, label_key]
+    columns:
+      - name: issue_key
+        data_tests:
+          - not_null
+          - relationships:
+              arguments:
+                to: ref('fct_issues')
+                field: issue_key
+      - name: label_key
+        data_tests:
+          - not_null
+          - relationships:
+              arguments:
+                to: ref('dim_labels')
+                field: label_key
+
+  - name: fct_daily_downloads
+    description: One row per package per day (PyPI 'without_mirrors' downloads).
+    columns:
+      - name: download_key
+        description: Surrogate key, hash of package + download_date.
+        data_tests: [not_null, unique]
+      - name: date_key
+        data_tests:
+          - not_null
+          - relationships:
+              arguments:
+                to: ref('dim_dates')
+                field: date_key
+      - name: download_count
+        data_tests:
+          - dbt_utils.expression_is_true:
+              arguments:
+                expression: ">= 0"

dbt/models/marts/bridge_issue_labels.sql

@@ -0,0 +1,27 @@
+-- Bridge table for the issue<->label many-to-many. Labels land in staging as a JSON
+-- array per issue; we explode them, recompute the issue surrogate key the same way as
+-- fct_issues, and resolve each label name to its label_key in dim_labels.
+-- Grain: one row per (issue, label).
+
+with issue_labels as (
+    select
+        repository,
+        issue_number,
+        unnest(json_extract(labels, '$[*]')) as label
+    from {{ ref('stg_github__issues') }}
+    where labels is not null
+),
+
+exploded as (
+    select
+        {{ dbt_utils.generate_surrogate_key(['repository', 'issue_number']) }} as issue_key,
+        label ->> '$.name' as label_name
+    from issue_labels
+    where (label ->> '$.name') is not null
+)
+
+select distinct
+    e.issue_key,
+    dl.label_key
+from exploded e
+inner join {{ ref('dim_labels') }} dl on e.label_name = dl.label_name

dbt/models/marts/dim_repositories.sql

@@ -21,7 +21,16 @@ with versioned as (
         -- unique even if a repo changes more than once on the same calendar day;
         -- valid_from/valid_to are kept at date grain for fact date-range joins.
         dbt_valid_from as version_ts,
-        dbt_valid_from::date as valid_from,
+        -- The earliest version opens at a past-infinity sentinel ('1900-01-01')
+        -- so historical facts that predate the first snapshot still resolve to
+        -- the first captured version instead of dropping.
+        case
+            when row_number() over (
+                partition by repository_id order by dbt_valid_from
+            ) = 1 then date '1900-01-01'
+            else dbt_valid_from::date
+        end as valid_from,
+        -- '9999-12-31' = future-infinity bound
         coalesce(dbt_valid_to::date, date '9999-12-31') as valid_to,
         dbt_valid_to is null as is_current
     from {{ ref('snap_repositories') }}

dbt/models/marts/fct_commits.sql

@@ -0,0 +1,25 @@
+-- Commit activity fact: one row per commit. Joins to dim_repositories via the SCD2
+-- half-open date range (event_date >= valid_from AND < valid_to) so each commit maps
+-- to the repository version that was current when it landed; to dim_contributors on
+-- author login; and to dim_dates via an inline YYYYMMDD key.
+
+with commits as (
+    select * from {{ ref('stg_github__commits') }}
+)
+
+select
+    {{ dbt_utils.generate_surrogate_key(['c.repository', 'c.commit_sha']) }} as commit_key,
+    r.repository_key,
+    co.contributor_key,
+    cast(strftime(c.committed_at, '%Y%m%d') as integer) as date_key,
+    c.commit_sha as commit_hash,
+    c.additions as lines_added,
+    c.deletions as lines_deleted,
+    c.committed_at
+from commits c
+left join {{ ref('dim_repositories') }} r
+    on c.repository = r.repository_name
+    and c.committed_at::date >= r.valid_from
+    and c.committed_at::date < r.valid_to
+left join {{ ref('dim_contributors') }} co
+    on c.author_login = co.username

dbt/models/marts/fct_daily_downloads.sql

@@ -0,0 +1,15 @@
+-- PyPI daily download fact: one row per package per day.
+-- Filtered to the 'without_mirrors' overall time-series category to avoid double
+-- counting 'with_mirrors' and to exclude the recent-endpoint 'last_*' aggregates.
+
+with downloads as (
+    select * from {{ ref('stg_pypi__downloads') }}
+    where category = 'without_mirrors'
+)
+
+select
+    {{ dbt_utils.generate_surrogate_key(['package', 'download_date']) }} as download_key,
+    package,
+    cast(strftime(download_date, '%Y%m%d') as integer) as date_key,
+    download_count
+from downloads

dbt/models/marts/fct_issues.sql

@@ -0,0 +1,24 @@
+-- Issue lifecycle fact: one row per issue. Repository resolved via the SCD2 half-open
+-- range on the issue open date; author via dim_contributors; opened/closed date keys
+-- are inline YYYYMMDD references to dim_dates (closed is nullable).
+
+with issues as (
+    select * from {{ ref('stg_github__issues') }}
+)
+
+select
+    {{ dbt_utils.generate_surrogate_key(['i.repository', 'i.issue_number']) }} as issue_key,
+    r.repository_key,
+    co.contributor_key as author_key,
+    cast(strftime(i.created_at, '%Y%m%d') as integer) as opened_date_key,
+    cast(strftime(i.closed_at, '%Y%m%d') as integer) as closed_date_key,
+    i.is_closed,
+    i.time_to_close_hours,
+    i.comment_count
+from issues i
+left join {{ ref('dim_repositories') }} r
+    on i.repository = r.repository_name
+    and i.created_at::date >= r.valid_from
+    and i.created_at::date < r.valid_to
+left join {{ ref('dim_contributors') }} co
+    on i.author_login = co.username

dbt/models/marts/fct_pull_requests.sql

@@ -0,0 +1,27 @@
+-- Pull request lifecycle fact: one row per PR. Repository resolved via the SCD2
+-- half-open range on the PR open date; author via dim_contributors; opened/merged
+-- date keys are inline YYYYMMDD references to dim_dates (merged is nullable).
+
+with pull_requests as (
+    select * from {{ ref('stg_github__pull_requests') }}
+)
+
+select
+    {{ dbt_utils.generate_surrogate_key(['p.repository', 'p.pr_number']) }} as pr_key,
+    r.repository_key,
+    co.contributor_key as author_key,
+    cast(strftime(p.created_at, '%Y%m%d') as integer) as opened_date_key,
+    cast(strftime(p.merged_at, '%Y%m%d') as integer) as merged_date_key,
+    p.merged_at is not null as is_merged,
+    p.time_to_merge_hours,
+    p.review_comments as review_count,
+    p.comment_count,
+    p.additions,
+    p.deletions
+from pull_requests p
+left join {{ ref('dim_repositories') }} r
+    on p.repository = r.repository_name
+    and p.created_at::date >= r.valid_from
+    and p.created_at::date < r.valid_to
+left join {{ ref('dim_contributors') }} co
+    on p.author_login = co.username