Report Generation + Resizable + API Changes

backend/alembic/versions/869cfd49ebd5_initial.py

@@ -27,7 +27,7 @@ def upgrade() -> None:
         sa.Column("oidc_id", sa.String(length=255), nullable=False),
         sa.Column(
             "updated_at",
-            sa.DateTime(),
+            sa.DateTime(timezone=True),
             server_default=sa.text("(CURRENT_TIMESTAMP)"),
             nullable=False,
         ),
@@ -42,13 +42,13 @@ def upgrade() -> None:
         sa.Column("creator_user_id", sa.String(length=26), nullable=False),
         sa.Column(
             "created_at",
-            sa.DateTime(),
+            sa.DateTime(timezone=True),
             server_default=sa.text("(CURRENT_TIMESTAMP)"),
             nullable=False,
         ),
         sa.Column(
             "updated_at",
-            sa.DateTime(),
+            sa.DateTime(timezone=True),
             server_default=sa.text("(CURRENT_TIMESTAMP)"),
             nullable=False,
         ),
@@ -65,11 +65,11 @@ def upgrade() -> None:
         sa.Column("user_id", sa.String(length=26), nullable=False),
         sa.Column(
             "started_at",
-            sa.DateTime(),
+            sa.DateTime(timezone=True),
             server_default=sa.text("(CURRENT_TIMESTAMP)"),
             nullable=False,
         ),
-        sa.Column("ended_at", sa.DateTime(), nullable=True),
+        sa.Column("ended_at", sa.DateTime(timezone=True), nullable=True),
         sa.ForeignKeyConstraint(
             ["project_id"],
             ["project.project_id"],
@@ -90,13 +90,13 @@ def upgrade() -> None:
         sa.Column("speaker", sa.String(length=100), nullable=True),
         sa.Column(
             "created_at",
-            sa.DateTime(),
+            sa.DateTime(timezone=True),
             server_default=sa.text("(CURRENT_TIMESTAMP)"),
             nullable=False,
         ),
         sa.Column(
             "updated_at",
-            sa.DateTime(),
+            sa.DateTime(timezone=True),
             server_default=sa.text("(CURRENT_TIMESTAMP)"),
             nullable=False,
         ),
@@ -115,15 +115,17 @@ def upgrade() -> None:
         sa.Column("analysis_id", sa.String(length=26), nullable=False),
         sa.Column("project_id", sa.String(length=26), nullable=False),
         sa.Column("text", sa.Text(), nullable=False),
+        sa.Column("category_code", sa.String(length=1), nullable=False),
         sa.Column("span", sa.Text(), nullable=True),
         sa.Column("transcript_span_id", sa.String(length=26), nullable=True),
         sa.Column("transcript_context_start", sa.String(length=26), nullable=False),
         sa.Column("transcript_context_end", sa.String(length=26), nullable=False),
         sa.Column("summary", sa.Text(), nullable=False),
         sa.Column("tag", sa.String(length=50), nullable=True),
-        sa.Column("time_tag_changed", sa.DateTime(), nullable=True),
+        sa.Column("time_tag_changed", sa.DateTime(timezone=True), nullable=True),
         sa.Column("was_asked", sa.Boolean(), nullable=True),
         sa.Column("asked_at_transcript_id", sa.String(length=26), nullable=True),
+        sa.Column("asked_at", sa.DateTime(timezone=True), nullable=True),
         sa.ForeignKeyConstraint(
             ["project_id"],
             ["project.project_id"],

backend/src/interview_helper/ai_analysis/ai_analysis.py

@@ -18,6 +18,10 @@
     ProjectId,
     TranscriptId,
 )
+from interview_helper.context_manager.question_categories import (
+    QUESTION_CATEGORIES,
+    normalize_question_category_code,
+)
 from langchain_openai import AzureChatOpenAI
 from langchain.tools import ToolRuntime, tool  # pyright: ignore[reportUnknownVariableType]
 from langchain.agents import create_agent  # pyright: ignore[reportUnknownVariableType]
@@ -40,6 +44,7 @@
 class Question(BaseModel):
     question: str
     grounding_span: str
+    category_code: str
 
 
 class Analysis(BaseModel):
@@ -55,7 +60,11 @@ class ProjectContext:
 class SimpleAnalyzer:
     """Simple LLM-based interview analyzer."""
 
-    SYSTEM_PROMPT: str = dedent("""\
+    CATEGORY_PROMPT_BLOCK: str = "\n".join(
+        [f"   - {code}: {label}" for code, label in QUESTION_CATEGORIES]
+    )
+
+    SYSTEM_PROMPT: str = dedent(f"""\
         ROLE: Interview Follow-Up Generator for SAR Profiles
 
         You will receive a chunk of transcript from an in-depth profile interview for a Search and Rescue operation.
@@ -79,8 +88,12 @@ class SimpleAnalyzer:
         7) Output: ONE to THREE questions, each with:
             - question (string)
             - grounding_span (short verbatim quote from the transcript)
+            - category_code (single letter in B-W from list below)
            As well as a brief summary of the entire situation so far, based on your knowledge.
 
+        Category codes:
+{CATEGORY_PROMPT_BLOCK}
+
         ALWAYS use the provided TOOLS to check for duplicates or 
         gather more context from the transcript history before finalizing your questions.
 
@@ -246,6 +259,7 @@ def clean_grounding_span(span: str) -> str:
                 AIQuestion(
                     question=q.question,
                     grounding_span=clean_grounding_span(q.grounding_span),
+                    category_code=normalize_question_category_code(q.category_code),
                 )
                 for q in analysis.questions
             ]

backend/src/interview_helper/config.py

@@ -50,7 +50,7 @@ def model_post_init(self, __context):
     bytes_per_sample: int = 2
 
     # AI Processing
-    process_transcript_every_secs: float = 60.0 * 2  # 2 minutes
+    process_transcript_every_secs: float = 60.0 * 4  # 4 minutes
     process_transcript_every_word_count: int = 100
 
     azure_api_endpoint: str = Field(alias="OPENAI_API_ENDPOINT")