Spaces:
Sleeping
Sleeping
Mark Cuizon commited on
Commit ·
f8d30ca
1
Parent(s): e400cff
fix: analysis progress UI, Gemini models, and Step 2/3/5 results
Browse filesShorten AI progress labels to stop duplicating full step prompts. Prefer gemini-2.5-pro/flash. Rebuild Step 2/3/5 narratives from real issues and economic forecast when LLM returns placeholders. Route Step 2 and 5 through dedicated renderers and hide empty Issues Source dumps.
- modules/results.py +162 -189
- utils/analysis_engine.py +154 -30
- utils/streamlit_runtime.py +32 -9
modules/results.py
CHANGED
|
@@ -7607,71 +7607,89 @@ def display_step2_issue_diagnosis(analysis_data):
|
|
| 7607 |
summary_text = analysis_data['summary']
|
| 7608 |
if isinstance(summary_text, str) and summary_text.strip():
|
| 7609 |
try:
|
| 7610 |
-
# Remove any raw dict dumps if the LLM leaked them into the summary
|
| 7611 |
summary_text = _strip_step5_raw_economic_blocks(summary_text)
|
| 7612 |
except Exception:
|
| 7613 |
pass
|
| 7614 |
-
|
| 7615 |
-
|
| 7616 |
-
|
| 7617 |
-
|
| 7618 |
-
|
| 7619 |
-
|
|
|
|
|
|
|
| 7620 |
|
| 7621 |
-
# Display identified issues
|
| 7622 |
-
|
|
|
|
| 7623 |
st.markdown(f"#### 🚨 {t('identified_issues', 'Identified Issues')}")
|
| 7624 |
-
issues = analysis_data['identified_issues']
|
| 7625 |
if isinstance(issues, list):
|
| 7626 |
-
|
| 7627 |
-
|
| 7628 |
-
|
| 7629 |
-
|
| 7630 |
-
st.warning(f"⚠️ Filtered out corrupted issue for parameter: {issue.get('parameter', 'Unknown')}")
|
| 7631 |
-
continue
|
| 7632 |
-
|
| 7633 |
-
# Create a formatted card for each issue
|
| 7634 |
-
parameter = issue.get('parameter', 'Unknown')
|
| 7635 |
-
current_value = issue.get('current_value', 'N/A')
|
| 7636 |
-
optimal_range = issue.get('optimal_range', 'N/A')
|
| 7637 |
-
status = issue.get('status', 'Unknown')
|
| 7638 |
-
severity = issue.get('severity', 'Unknown')
|
| 7639 |
-
description = issue.get('description', '')
|
| 7640 |
-
|
| 7641 |
st.markdown(
|
| 7642 |
-
f'<div style="margin-bottom:
|
| 7643 |
-
f'<
|
| 7644 |
-
f'<div style="display: grid; grid-template-columns: 1fr 1fr; gap: 10px; margin-bottom: 10px;">'
|
| 7645 |
-
f'<div><strong>Current Value:</strong> {current_value}</div>'
|
| 7646 |
-
f'<div><strong>Optimal Range:</strong> {optimal_range}</div>'
|
| 7647 |
-
f'<div><strong>Status:</strong> {status}</div>'
|
| 7648 |
-
f'<div><strong>Severity:</strong> {severity}</div>'
|
| 7649 |
-
f'</div>'
|
| 7650 |
-
f'{f"<div><strong>Description:</strong> {description}</div>" if description else ""}'
|
| 7651 |
f'</div>',
|
| 7652 |
unsafe_allow_html=True
|
| 7653 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7654 |
else:
|
| 7655 |
-
|
| 7656 |
-
|
| 7657 |
-
|
| 7658 |
-
f'<p style="margin: 0; font-size: 15px; color: #2c3e50;"><strong>Issue {i}:</strong> {sanitize_persona_and_enforce_article(str(issue))}</p>'
|
| 7659 |
-
f'</div>',
|
| 7660 |
-
unsafe_allow_html=True
|
| 7661 |
-
)
|
| 7662 |
-
|
| 7663 |
-
# Display visualizations
|
| 7664 |
-
if 'visualizations' in analysis_data and analysis_data['visualizations']:
|
| 7665 |
-
st.markdown("#### 📊 Visualizations")
|
| 7666 |
-
try:
|
| 7667 |
-
visualizations = analysis_data['visualizations']
|
| 7668 |
-
if isinstance(visualizations, list):
|
| 7669 |
-
for i, viz in enumerate(visualizations, 1):
|
| 7670 |
-
if isinstance(viz, dict) and 'type' in viz:
|
| 7671 |
-
display_visualization(viz, i, 2)
|
| 7672 |
-
except Exception as e:
|
| 7673 |
-
logger.error(f"Error displaying visualizations: {e}")
|
| 7674 |
-
st.error("Error displaying visualizations")
|
| 7675 |
|
| 7676 |
# Display tables
|
| 7677 |
if 'tables' in analysis_data and analysis_data['tables']:
|
|
@@ -7684,73 +7702,40 @@ def display_step2_issue_diagnosis(analysis_data):
|
|
| 7684 |
display_table(table_data, f"Table {i}")
|
| 7685 |
except Exception as e:
|
| 7686 |
logger.error(f"Error displaying tables: {e}")
|
| 7687 |
-
st.error("Error displaying tables")
|
| 7688 |
|
| 7689 |
-
#
|
| 7690 |
-
# Check formatted_analysis first, then detailed_analysis
|
| 7691 |
analysis_text = None
|
| 7692 |
if 'formatted_analysis' in analysis_data and analysis_data['formatted_analysis']:
|
| 7693 |
analysis_text = analysis_data['formatted_analysis']
|
| 7694 |
elif 'detailed_analysis' in analysis_data and analysis_data['detailed_analysis']:
|
| 7695 |
analysis_text = analysis_data['detailed_analysis']
|
| 7696 |
-
|
| 7697 |
-
if analysis_text
|
|
|
|
|
|
|
| 7698 |
st.markdown(f"#### 📋 {t('detailed_analysis', 'Detailed Analysis')}")
|
| 7699 |
-
|
| 7700 |
-
if isinstance(analysis_text, dict):
|
| 7701 |
analysis_text = str(analysis_text)
|
| 7702 |
-
elif not isinstance(analysis_text, str):
|
| 7703 |
-
analysis_text = str(analysis_text) if analysis_text is not None else ""
|
| 7704 |
-
|
| 7705 |
-
# If the LLM included a prefixed "Formatted Analysis:" section, prefer its content
|
| 7706 |
try:
|
| 7707 |
import re
|
| 7708 |
-
|
| 7709 |
-
|
| 7710 |
-
analysis_text = formatted_block.group(1).strip()
|
| 7711 |
-
except Exception:
|
| 7712 |
-
pass
|
| 7713 |
-
|
| 7714 |
-
# Clean HTML-like tags (e.g., <br>, <br/>, <br />) and convert to line breaks
|
| 7715 |
-
try:
|
| 7716 |
-
import re as _re_html
|
| 7717 |
-
# Replace <br>, <br/>, <br /> with actual line breaks
|
| 7718 |
-
analysis_text = _re_html.sub(r'<br\s*/?>', '\n\n', analysis_text, flags=_re_html.IGNORECASE)
|
| 7719 |
-
# Remove other common HTML tags that might leak through
|
| 7720 |
-
analysis_text = _re_html.sub(r'</?[a-z]+[^>]*>', '', analysis_text, flags=_re_html.IGNORECASE)
|
| 7721 |
except Exception:
|
| 7722 |
pass
|
| 7723 |
-
|
| 7724 |
-
# Unescape common escape sequences that sometimes leak from LLM
|
| 7725 |
-
try:
|
| 7726 |
-
analysis_text = analysis_text.replace('\\n', '\n').replace('\\t', '\t').replace('\\"', '"')
|
| 7727 |
-
except Exception:
|
| 7728 |
-
pass
|
| 7729 |
-
|
| 7730 |
-
# Sanitize persona and enforce neutral tone before rendering
|
| 7731 |
analysis_text = sanitize_persona_and_enforce_article(analysis_text)
|
| 7732 |
-
|
| 7733 |
-
# Extract and render markdown tables first
|
| 7734 |
-
analysis_text = _extract_and_render_markdown_tables(analysis_text)
|
| 7735 |
-
|
| 7736 |
-
# Normalize markdown formatting
|
| 7737 |
-
analysis_text = normalize_markdown_block_for_step3(analysis_text)
|
| 7738 |
-
|
| 7739 |
-
# Split into paragraphs and display
|
| 7740 |
-
paragraphs = analysis_text.split('\n\n') if '\n\n' in analysis_text else [analysis_text]
|
| 7741 |
-
for paragraph in paragraphs:
|
| 7742 |
if paragraph.strip():
|
| 7743 |
-
|
| 7744 |
-
if any(token in paragraph for token in ['\n- ', '\n* ', '\n1.', '\n2.', '### ', '#### ', '**']):
|
| 7745 |
st.markdown(paragraph)
|
| 7746 |
else:
|
| 7747 |
st.markdown(
|
| 7748 |
-
f'<div style="margin-bottom: 18px; padding: 15px; background: linear-gradient(135deg, #ffffff, #f8f9fa); border: 1px solid #e9ecef; border-radius: 8px;
|
| 7749 |
f'<p style="margin: 0; line-height: 1.8; font-size: 16px; color: #2c3e50;">{paragraph.strip()}</p>'
|
| 7750 |
f'</div>',
|
| 7751 |
unsafe_allow_html=True
|
| 7752 |
)
|
| 7753 |
-
|
|
|
|
| 7754 |
|
| 7755 |
def display_step5_economic_forecast(analysis_data):
|
| 7756 |
"""Display Step 5: Economic Impact Forecast content with consistent formatting"""
|
|
@@ -9552,34 +9537,49 @@ def display_step_block(step_result, step_number, step_title):
|
|
| 9552 |
display_enhanced_step_result(step_result, step_number)
|
| 9553 |
|
| 9554 |
def display_enhanced_step_result(step_result, step_number):
|
| 9555 |
-
"""Display enhanced step results with proper structure and formatting for non-technical users
|
| 9556 |
-
|
|
|
|
|
|
|
|
|
|
| 9557 |
if not isinstance(step_result, dict):
|
| 9558 |
logger.error(f"Step {step_number} step_result is not a dictionary: {type(step_result)}")
|
| 9559 |
st.error(f"❌ Error: Step {step_number} data is not in the expected format")
|
| 9560 |
return
|
| 9561 |
-
|
| 9562 |
analysis_data = step_result
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9563 |
|
| 9564 |
-
# Debug: Log the structure of analysis_data
|
| 9565 |
logger.info(f"Step {step_number} analysis_data keys: {list(analysis_data.keys()) if analysis_data else 'None'}")
|
| 9566 |
-
|
| 9567 |
-
|
| 9568 |
-
if 'detailed_analysis' in analysis_data:
|
| 9569 |
-
logger.info(f"Step {step_number} detailed_analysis type: {type(analysis_data['detailed_analysis'])}, length: {len(str(analysis_data['detailed_analysis'])) if analysis_data['detailed_analysis'] else 0}")
|
| 9570 |
-
|
| 9571 |
-
# Special handling for STEP 1 - Data Analysis
|
| 9572 |
if step_number == 1:
|
| 9573 |
display_step1_data_analysis(analysis_data)
|
| 9574 |
return
|
| 9575 |
-
|
| 9576 |
-
|
| 9577 |
-
|
| 9578 |
-
# Special handling for STEP 3 - Solution Recommendations
|
| 9579 |
if step_number == 3:
|
| 9580 |
display_step3_solution_recommendations(analysis_data)
|
| 9581 |
return
|
| 9582 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9583 |
# Alias-mapping: normalize common capitalized keys from LLM output
|
| 9584 |
try:
|
| 9585 |
alias_map = {
|
|
@@ -9600,7 +9600,6 @@ def display_enhanced_step_result(step_result, step_number):
|
|
| 9600 |
for k, v in list(analysis_data.items()):
|
| 9601 |
if k in alias_map and alias_map[k] not in analysis_data:
|
| 9602 |
analysis_data[alias_map[k]] = v
|
| 9603 |
-
# Remove original capitalized keys to prevent raw dict leakage in other_fields
|
| 9604 |
for original_key in list(analysis_data.keys()):
|
| 9605 |
if original_key in alias_map:
|
| 9606 |
try:
|
|
@@ -9610,11 +9609,13 @@ def display_enhanced_step_result(step_result, step_number):
|
|
| 9610 |
except Exception:
|
| 9611 |
pass
|
| 9612 |
|
| 9613 |
-
# 1. SUMMARY SECTION
|
| 9614 |
if 'summary' in analysis_data and analysis_data['summary']:
|
| 9615 |
-
st.markdown(f"### 📋 {t('summary', 'Summary')}")
|
| 9616 |
summary_text = analysis_data['summary']
|
| 9617 |
-
if isinstance(summary_text, str) and summary_text.strip()
|
|
|
|
|
|
|
|
|
|
| 9618 |
st.markdown(
|
| 9619 |
f'<div style="margin-bottom: 20px; padding: 15px; background: linear-gradient(135deg, #e8f5e8, #ffffff); border-left: 4px solid #28a745; border-radius: 8px; box-shadow: 0 2px 8px rgba(0,0,0,0.1);">'
|
| 9620 |
f'<p style="margin: 0; font-size: 16px; line-height: 1.6; color: #2c3e50;">{summary_text.strip()}</p>'
|
|
@@ -9622,82 +9623,41 @@ def display_enhanced_step_result(step_result, step_number):
|
|
| 9622 |
unsafe_allow_html=True
|
| 9623 |
)
|
| 9624 |
st.markdown("")
|
| 9625 |
-
|
| 9626 |
-
#
|
| 9627 |
-
# Key findings are now consolidated and displayed only after Executive Summary
|
| 9628 |
-
|
| 9629 |
-
# 3. DETAILED ANALYSIS SECTION - Show if available (with filtering for all steps)
|
| 9630 |
if 'detailed_analysis' in analysis_data and analysis_data['detailed_analysis']:
|
| 9631 |
-
st.markdown(f"### 📋 {t('detailed_analysis', 'Detailed Analysis')}")
|
| 9632 |
detailed_text = analysis_data['detailed_analysis']
|
| 9633 |
-
|
| 9634 |
-
# Ensure detailed_text is a string
|
| 9635 |
if isinstance(detailed_text, dict):
|
| 9636 |
detailed_text = str(detailed_text)
|
| 9637 |
elif not isinstance(detailed_text, str):
|
| 9638 |
-
detailed_text = str(detailed_text) if detailed_text is not None else "
|
| 9639 |
-
|
| 9640 |
-
|
| 9641 |
-
|
| 9642 |
-
|
| 9643 |
-
|
| 9644 |
-
|
| 9645 |
-
|
| 9646 |
-
|
| 9647 |
-
|
| 9648 |
-
|
| 9649 |
-
|
| 9650 |
-
|
| 9651 |
-
|
| 9652 |
-
|
| 9653 |
-
|
| 9654 |
-
|
| 9655 |
-
|
| 9656 |
-
|
| 9657 |
-
|
| 9658 |
-
|
| 9659 |
-
|
| 9660 |
-
|
| 9661 |
-
for paragraph in paragraphs:
|
| 9662 |
-
if isinstance(paragraph, str) and paragraph.strip():
|
| 9663 |
-
# Skip empty paragraphs
|
| 9664 |
-
if paragraph.strip() == '':
|
| 9665 |
-
continue
|
| 9666 |
-
|
| 9667 |
-
# Check if this paragraph contains a table (already processed)
|
| 9668 |
-
if '<table' in paragraph and '</table>' in paragraph:
|
| 9669 |
-
# This is an HTML table, render it directly
|
| 9670 |
-
st.markdown(paragraph, unsafe_allow_html=True)
|
| 9671 |
-
else:
|
| 9672 |
-
# Regular text paragraph
|
| 9673 |
-
st.markdown(
|
| 9674 |
-
f'<div style="margin-bottom: 18px; padding: 15px; background: linear-gradient(135deg, #ffffff, #f8f9fa); border: 1px solid #e9ecef; border-radius: 8px; box-shadow: 0 2px 4px rgba(0,0,0,0.05);">'
|
| 9675 |
-
f'<p style="margin: 0; line-height: 1.8; font-size: 16px; color: #2c3e50;">{sanitize_persona_and_enforce_article(paragraph.strip())}</p>'
|
| 9676 |
-
f'</div>',
|
| 9677 |
-
unsafe_allow_html=True
|
| 9678 |
-
)
|
| 9679 |
-
else:
|
| 9680 |
-
# If all content was filtered out, don't display anything
|
| 9681 |
-
pass
|
| 9682 |
-
st.markdown("")
|
| 9683 |
-
|
| 9684 |
-
# 3.5. PLANTATION VALUES VS REFERENCE RANGES SECTION - REMOVED as requested
|
| 9685 |
-
# This section has been removed to eliminate the "Visual Comparison: Plantation vs. Malaysian Reference Ranges" display
|
| 9686 |
-
|
| 9687 |
-
# 3.6. SOIL ISSUES SECTION - Disabled to prevent raw or malformed 'Soil Issues' output
|
| 9688 |
-
# Intentionally no-op: we do not render analysis_data['soil_issues'] at all
|
| 9689 |
-
|
| 9690 |
-
st.markdown("")
|
| 9691 |
|
| 9692 |
-
#
|
| 9693 |
if 'scenarios' in analysis_data and analysis_data['scenarios']:
|
| 9694 |
scenarios = analysis_data['scenarios']
|
| 9695 |
-
|
| 9696 |
st.markdown(f"### 📈 {t('investment_scenarios', 'Investment Scenarios')}")
|
| 9697 |
-
|
| 9698 |
-
# Handle different formats of scenarios data
|
| 9699 |
if isinstance(scenarios, dict):
|
| 9700 |
-
# Display each scenario in a formatted card
|
| 9701 |
for scenario_key, scenario_data in scenarios.items():
|
| 9702 |
if isinstance(scenario_data, dict):
|
| 9703 |
display_formatted_scenario(scenario_key, scenario_data)
|
|
@@ -9935,12 +9895,19 @@ def display_enhanced_step_result(step_result, step_number):
|
|
| 9935 |
'tables', 'interpretations', 'data_quality',
|
| 9936 |
'specific_recommendations', 'format_analysis', 'data_format_recommendations',
|
| 9937 |
'raw_llm_output', 'raw_output', 'raw_llm',
|
| 9938 |
-
'scenarios', 'assumptions'
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9939 |
])
|
| 9940 |
excluded_keys.update([
|
| 9941 |
'Key Findings', 'Specific Recommendations', 'Tables', 'Interpretations', 'Visualizations',
|
| 9942 |
'Yield Forecast', 'Format Analysis', 'Data Format Recommendations', 'Plantation Values vs. Malaysian Reference Ranges',
|
| 9943 |
-
'Soil Issues', 'Issues Source'
|
|
|
|
| 9944 |
])
|
| 9945 |
# Also exclude capitalized variants that may appear from LLM outputs
|
| 9946 |
excluded_keys.update([
|
|
@@ -9952,8 +9919,14 @@ def display_enhanced_step_result(step_result, step_number):
|
|
| 9952 |
'raw_llm_output', 'raw_output', 'raw_llm', 'deterministic', 'llm_output',
|
| 9953 |
'Item 0', 'Item 1', 'Item 2', 'Item 3', 'Item 4', 'Item 5', 'Item 6', 'Item 7', 'Item 8', 'Item 9'
|
| 9954 |
])
|
| 9955 |
-
other_fields = [
|
| 9956 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9957 |
has_key_findings = bool(analysis_data.get('key_findings'))
|
| 9958 |
if has_key_findings or other_fields:
|
| 9959 |
st.markdown(f"### 📊 {t('additional_analysis_results', 'Additional Analysis Results')}")
|
|
|
|
| 7607 |
summary_text = analysis_data['summary']
|
| 7608 |
if isinstance(summary_text, str) and summary_text.strip():
|
| 7609 |
try:
|
|
|
|
| 7610 |
summary_text = _strip_step5_raw_economic_blocks(summary_text)
|
| 7611 |
except Exception:
|
| 7612 |
pass
|
| 7613 |
+
# Hide useless placeholder summaries
|
| 7614 |
+
if summary_text.strip().lower() not in ('analysis completed', 'detailed analysis not available'):
|
| 7615 |
+
st.markdown(
|
| 7616 |
+
f'<div style="margin-bottom: 20px; padding: 15px; background: linear-gradient(135deg, #e8f5e8, #ffffff); border-left: 4px solid #28a745; border-radius: 8px; box-shadow: 0 2px 8px rgba(0,0,0,0.1);">'
|
| 7617 |
+
f'<p style="margin: 0; font-size: 16px; line-height: 1.6; color: #2c3e50;">{sanitize_persona_and_enforce_article(summary_text.strip())}</p>'
|
| 7618 |
+
f'</div>',
|
| 7619 |
+
unsafe_allow_html=True
|
| 7620 |
+
)
|
| 7621 |
|
| 7622 |
+
# Display identified issues (supports both LLM and deterministic shapes)
|
| 7623 |
+
issues = analysis_data.get('identified_issues') or []
|
| 7624 |
+
if issues:
|
| 7625 |
st.markdown(f"#### 🚨 {t('identified_issues', 'Identified Issues')}")
|
|
|
|
| 7626 |
if isinstance(issues, list):
|
| 7627 |
+
issue_num = 0
|
| 7628 |
+
for issue in issues:
|
| 7629 |
+
if not isinstance(issue, dict):
|
| 7630 |
+
issue_num += 1
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7631 |
st.markdown(
|
| 7632 |
+
f'<div style="margin-bottom: 15px; padding: 12px; background: linear-gradient(135deg, #fff5f5, #ffffff); border-left: 4px solid #dc3545; border-radius: 6px;">'
|
| 7633 |
+
f'<p style="margin: 0; font-size: 15px; color: #2c3e50;"><strong>Issue {issue_num}:</strong> {sanitize_persona_and_enforce_article(str(issue))}</p>'
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7634 |
f'</div>',
|
| 7635 |
unsafe_allow_html=True
|
| 7636 |
)
|
| 7637 |
+
continue
|
| 7638 |
+
if issue.get('type') == 'Summary' or issue.get('parameter') == 'Summary':
|
| 7639 |
+
continue
|
| 7640 |
+
try:
|
| 7641 |
+
if _is_corrupted_soil_issue(issue):
|
| 7642 |
+
continue
|
| 7643 |
+
except Exception:
|
| 7644 |
+
pass
|
| 7645 |
+
|
| 7646 |
+
issue_num += 1
|
| 7647 |
+
parameter = issue.get('parameter', 'Unknown')
|
| 7648 |
+
itype = issue.get('type') or issue.get('issue_type') or 'Issue'
|
| 7649 |
+
current_value = issue.get('current_value', issue.get('value', 'N/A'))
|
| 7650 |
+
optimal_range = issue.get('optimal_range', issue.get('threshold', 'See MPOB standards'))
|
| 7651 |
+
severity = issue.get('severity', issue.get('status', 'Unknown'))
|
| 7652 |
+
cause = issue.get('cause', '')
|
| 7653 |
+
impact = issue.get('impact', '')
|
| 7654 |
+
recommendation = issue.get('recommendation', '')
|
| 7655 |
+
description = issue.get('description', '') or issue.get('issue_description', '')
|
| 7656 |
+
|
| 7657 |
+
detail_bits = []
|
| 7658 |
+
if cause:
|
| 7659 |
+
detail_bits.append(f"<div><strong>Cause:</strong> {cause}</div>")
|
| 7660 |
+
if impact:
|
| 7661 |
+
detail_bits.append(f"<div><strong>Impact:</strong> {impact}</div>")
|
| 7662 |
+
if recommendation:
|
| 7663 |
+
detail_bits.append(f"<div><strong>Recommended action:</strong> {recommendation}</div>")
|
| 7664 |
+
if description:
|
| 7665 |
+
detail_bits.append(f"<div><strong>Details:</strong> {description}</div>")
|
| 7666 |
+
details_html = "".join(detail_bits)
|
| 7667 |
+
|
| 7668 |
+
st.markdown(
|
| 7669 |
+
f'<div style="margin-bottom: 20px; padding: 15px; background: linear-gradient(135deg, #fff5f5, #ffffff); border-left: 4px solid #dc3545; border-radius: 8px; box-shadow: 0 2px 8px rgba(0,0,0,0.1);">'
|
| 7670 |
+
f'<h4 style="margin: 0 0 10px 0; color: #dc3545; font-size: 18px;">⚠️ Issue {issue_num}: {parameter} <span style="font-size:14px;color:#666;">({itype})</span></h4>'
|
| 7671 |
+
f'<div style="display: grid; grid-template-columns: 1fr 1fr; gap: 10px; margin-bottom: 10px;">'
|
| 7672 |
+
f'<div><strong>Current Value:</strong> {current_value}</div>'
|
| 7673 |
+
f'<div><strong>Reference:</strong> {optimal_range}</div>'
|
| 7674 |
+
f'<div><strong>Severity:</strong> {severity}</div>'
|
| 7675 |
+
f'</div>'
|
| 7676 |
+
f'{details_html}'
|
| 7677 |
+
f'</div>',
|
| 7678 |
+
unsafe_allow_html=True
|
| 7679 |
+
)
|
| 7680 |
+
|
| 7681 |
+
# Soil / leaf issue lists (from standards comparator) when present as strings
|
| 7682 |
+
for label, key in (("Soil Issues", "soil_issues"), ("Leaf Issues", "leaf_issues")):
|
| 7683 |
+
vals = analysis_data.get(key) or []
|
| 7684 |
+
if isinstance(vals, list) and vals and not issues:
|
| 7685 |
+
st.markdown(f"#### 🌱 {label}" if "Soil" in label else f"#### 🍃 {label}")
|
| 7686 |
+
for i, item in enumerate(vals, 1):
|
| 7687 |
+
if isinstance(item, dict):
|
| 7688 |
+
text = item.get('description') or item.get('issue') or item.get('message') or str(item)
|
| 7689 |
else:
|
| 7690 |
+
text = str(item)
|
| 7691 |
+
if text.strip():
|
| 7692 |
+
st.markdown(f"{i}. {sanitize_persona_and_enforce_article(text.strip())}")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7693 |
|
| 7694 |
# Display tables
|
| 7695 |
if 'tables' in analysis_data and analysis_data['tables']:
|
|
|
|
| 7702 |
display_table(table_data, f"Table {i}")
|
| 7703 |
except Exception as e:
|
| 7704 |
logger.error(f"Error displaying tables: {e}")
|
|
|
|
| 7705 |
|
| 7706 |
+
# Detailed analysis narrative
|
|
|
|
| 7707 |
analysis_text = None
|
| 7708 |
if 'formatted_analysis' in analysis_data and analysis_data['formatted_analysis']:
|
| 7709 |
analysis_text = analysis_data['formatted_analysis']
|
| 7710 |
elif 'detailed_analysis' in analysis_data and analysis_data['detailed_analysis']:
|
| 7711 |
analysis_text = analysis_data['detailed_analysis']
|
| 7712 |
+
|
| 7713 |
+
if analysis_text and str(analysis_text).strip().lower() not in (
|
| 7714 |
+
'detailed analysis not available', 'analysis completed', ''
|
| 7715 |
+
):
|
| 7716 |
st.markdown(f"#### 📋 {t('detailed_analysis', 'Detailed Analysis')}")
|
| 7717 |
+
if not isinstance(analysis_text, str):
|
|
|
|
| 7718 |
analysis_text = str(analysis_text)
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7719 |
try:
|
| 7720 |
import re
|
| 7721 |
+
analysis_text = re.sub(r'<br\s*/?>', '\n\n', analysis_text, flags=re.I)
|
| 7722 |
+
analysis_text = re.sub(r'</?[a-z]+[^>]*>', '', analysis_text, flags=re.I)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7723 |
except Exception:
|
| 7724 |
pass
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7725 |
analysis_text = sanitize_persona_and_enforce_article(analysis_text)
|
| 7726 |
+
for paragraph in (analysis_text.split('\n\n') if '\n\n' in analysis_text else [analysis_text]):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7727 |
if paragraph.strip():
|
| 7728 |
+
if any(token in paragraph for token in ['\n- ', '\n* ', '### ', '**']):
|
|
|
|
| 7729 |
st.markdown(paragraph)
|
| 7730 |
else:
|
| 7731 |
st.markdown(
|
| 7732 |
+
f'<div style="margin-bottom: 18px; padding: 15px; background: linear-gradient(135deg, #ffffff, #f8f9fa); border: 1px solid #e9ecef; border-radius: 8px;">'
|
| 7733 |
f'<p style="margin: 0; line-height: 1.8; font-size: 16px; color: #2c3e50;">{paragraph.strip()}</p>'
|
| 7734 |
f'</div>',
|
| 7735 |
unsafe_allow_html=True
|
| 7736 |
)
|
| 7737 |
+
elif not issues:
|
| 7738 |
+
st.info("No critical agronomic issues were flagged from the current soil and leaf averages.")
|
| 7739 |
|
| 7740 |
def display_step5_economic_forecast(analysis_data):
|
| 7741 |
"""Display Step 5: Economic Impact Forecast content with consistent formatting"""
|
|
|
|
| 9537 |
display_enhanced_step_result(step_result, step_number)
|
| 9538 |
|
| 9539 |
def display_enhanced_step_result(step_result, step_number):
|
| 9540 |
+
"""Display enhanced step results with proper structure and formatting for non-technical users.
|
| 9541 |
+
|
| 9542 |
+
NOTE: This is the active implementation (defined last). Step-specific renderers must be
|
| 9543 |
+
routed here so Step 2/5 content is not lost to the generic "Additional Analysis Results" dump.
|
| 9544 |
+
"""
|
| 9545 |
if not isinstance(step_result, dict):
|
| 9546 |
logger.error(f"Step {step_number} step_result is not a dictionary: {type(step_result)}")
|
| 9547 |
st.error(f"❌ Error: Step {step_number} data is not in the expected format")
|
| 9548 |
return
|
| 9549 |
+
|
| 9550 |
analysis_data = step_result
|
| 9551 |
+
try:
|
| 9552 |
+
step_number = int(step_number)
|
| 9553 |
+
except Exception:
|
| 9554 |
+
pass
|
| 9555 |
|
|
|
|
| 9556 |
logger.info(f"Step {step_number} analysis_data keys: {list(analysis_data.keys()) if analysis_data else 'None'}")
|
| 9557 |
+
|
| 9558 |
+
# Route specialized steps first
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9559 |
if step_number == 1:
|
| 9560 |
display_step1_data_analysis(analysis_data)
|
| 9561 |
return
|
| 9562 |
+
if step_number == 2:
|
| 9563 |
+
display_step2_issue_diagnosis(analysis_data)
|
| 9564 |
+
return
|
|
|
|
| 9565 |
if step_number == 3:
|
| 9566 |
display_step3_solution_recommendations(analysis_data)
|
| 9567 |
return
|
| 9568 |
+
if step_number == 4:
|
| 9569 |
+
try:
|
| 9570 |
+
display_regenerative_agriculture_content(analysis_data)
|
| 9571 |
+
return
|
| 9572 |
+
except Exception:
|
| 9573 |
+
pass
|
| 9574 |
+
if step_number == 5:
|
| 9575 |
+
try:
|
| 9576 |
+
display_step5_economic_forecast(analysis_data)
|
| 9577 |
+
return
|
| 9578 |
+
except Exception as e:
|
| 9579 |
+
logger.error(f"Step 5 special handling failed: {e}", exc_info=True)
|
| 9580 |
+
st.error(f"Error displaying Step 5 economic forecast: {e}")
|
| 9581 |
+
# fall through to generic
|
| 9582 |
+
|
| 9583 |
# Alias-mapping: normalize common capitalized keys from LLM output
|
| 9584 |
try:
|
| 9585 |
alias_map = {
|
|
|
|
| 9600 |
for k, v in list(analysis_data.items()):
|
| 9601 |
if k in alias_map and alias_map[k] not in analysis_data:
|
| 9602 |
analysis_data[alias_map[k]] = v
|
|
|
|
| 9603 |
for original_key in list(analysis_data.keys()):
|
| 9604 |
if original_key in alias_map:
|
| 9605 |
try:
|
|
|
|
| 9609 |
except Exception:
|
| 9610 |
pass
|
| 9611 |
|
| 9612 |
+
# 1. SUMMARY SECTION
|
| 9613 |
if 'summary' in analysis_data and analysis_data['summary']:
|
|
|
|
| 9614 |
summary_text = analysis_data['summary']
|
| 9615 |
+
if isinstance(summary_text, str) and summary_text.strip().lower() not in (
|
| 9616 |
+
'analysis completed', 'detailed analysis not available', ''
|
| 9617 |
+
):
|
| 9618 |
+
st.markdown(f"### 📋 {t('summary', 'Summary')}")
|
| 9619 |
st.markdown(
|
| 9620 |
f'<div style="margin-bottom: 20px; padding: 15px; background: linear-gradient(135deg, #e8f5e8, #ffffff); border-left: 4px solid #28a745; border-radius: 8px; box-shadow: 0 2px 8px rgba(0,0,0,0.1);">'
|
| 9621 |
f'<p style="margin: 0; font-size: 16px; line-height: 1.6; color: #2c3e50;">{summary_text.strip()}</p>'
|
|
|
|
| 9623 |
unsafe_allow_html=True
|
| 9624 |
)
|
| 9625 |
st.markdown("")
|
| 9626 |
+
|
| 9627 |
+
# 3. DETAILED ANALYSIS SECTION
|
|
|
|
|
|
|
|
|
|
| 9628 |
if 'detailed_analysis' in analysis_data and analysis_data['detailed_analysis']:
|
|
|
|
| 9629 |
detailed_text = analysis_data['detailed_analysis']
|
|
|
|
|
|
|
| 9630 |
if isinstance(detailed_text, dict):
|
| 9631 |
detailed_text = str(detailed_text)
|
| 9632 |
elif not isinstance(detailed_text, str):
|
| 9633 |
+
detailed_text = str(detailed_text) if detailed_text is not None else ""
|
| 9634 |
+
if detailed_text.strip().lower() not in ('detailed analysis not available', 'analysis completed', ''):
|
| 9635 |
+
st.markdown(f"### 📋 {t('detailed_analysis', 'Detailed Analysis')}")
|
| 9636 |
+
detailed_text = sanitize_persona_and_enforce_article(detailed_text)
|
| 9637 |
+
import re
|
| 9638 |
+
detailed_text = re.sub(r'!\[.*?\]\(https://quickchart\.io.*?\)', '', detailed_text, flags=re.DOTALL)
|
| 9639 |
+
filtered_detailed_text = filter_known_sections_from_text(detailed_text)
|
| 9640 |
+
if filtered_detailed_text.strip() and filtered_detailed_text != "Content filtered to prevent raw LLM output display.":
|
| 9641 |
+
processed_text = process_html_tables(filtered_detailed_text)
|
| 9642 |
+
paragraphs = processed_text.split('\n\n') if '\n\n' in processed_text else [processed_text]
|
| 9643 |
+
for paragraph in paragraphs:
|
| 9644 |
+
if isinstance(paragraph, str) and paragraph.strip():
|
| 9645 |
+
if '<table' in paragraph and '</table>' in paragraph:
|
| 9646 |
+
st.markdown(paragraph, unsafe_allow_html=True)
|
| 9647 |
+
else:
|
| 9648 |
+
st.markdown(
|
| 9649 |
+
f'<div style="margin-bottom: 18px; padding: 15px; background: linear-gradient(135deg, #ffffff, #f8f9fa); border: 1px solid #e9ecef; border-radius: 8px; box-shadow: 0 2px 4px rgba(0,0,0,0.05);">'
|
| 9650 |
+
f'<p style="margin: 0; line-height: 1.8; font-size: 16px; color: #2c3e50;">{sanitize_persona_and_enforce_article(paragraph.strip())}</p>'
|
| 9651 |
+
f'</div>',
|
| 9652 |
+
unsafe_allow_html=True
|
| 9653 |
+
)
|
| 9654 |
+
st.markdown("")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9655 |
|
| 9656 |
+
# Economic scenarios if present (non-step-5 path)
|
| 9657 |
if 'scenarios' in analysis_data and analysis_data['scenarios']:
|
| 9658 |
scenarios = analysis_data['scenarios']
|
|
|
|
| 9659 |
st.markdown(f"### 📈 {t('investment_scenarios', 'Investment Scenarios')}")
|
|
|
|
|
|
|
| 9660 |
if isinstance(scenarios, dict):
|
|
|
|
| 9661 |
for scenario_key, scenario_data in scenarios.items():
|
| 9662 |
if isinstance(scenario_data, dict):
|
| 9663 |
display_formatted_scenario(scenario_key, scenario_data)
|
|
|
|
| 9895 |
'tables', 'interpretations', 'data_quality',
|
| 9896 |
'specific_recommendations', 'format_analysis', 'data_format_recommendations',
|
| 9897 |
'raw_llm_output', 'raw_output', 'raw_llm',
|
| 9898 |
+
'scenarios', 'assumptions',
|
| 9899 |
+
# Internal / already-rendered diagnosis fields (do not dump as "Analysis:")
|
| 9900 |
+
'identified_issues', 'soil_issues', 'leaf_issues', 'issues_source',
|
| 9901 |
+
'total_issues', 'economic_forecast', 'economic_forecast_source',
|
| 9902 |
+
'solution_options', 'nutrient_comparisons', 'visualizations_source',
|
| 9903 |
+
'success', 'processing_time', 'error_details', 'confidence_level', 'analysis',
|
| 9904 |
+
'regenerative_practices',
|
| 9905 |
])
|
| 9906 |
excluded_keys.update([
|
| 9907 |
'Key Findings', 'Specific Recommendations', 'Tables', 'Interpretations', 'Visualizations',
|
| 9908 |
'Yield Forecast', 'Format Analysis', 'Data Format Recommendations', 'Plantation Values vs. Malaysian Reference Ranges',
|
| 9909 |
+
'Soil Issues', 'Issues Source', 'Identified Issues', 'Soil Parameters Diagnosis',
|
| 9910 |
+
'Leaf Nutrient Diagnosis', 'Leaf Issues',
|
| 9911 |
])
|
| 9912 |
# Also exclude capitalized variants that may appear from LLM outputs
|
| 9913 |
excluded_keys.update([
|
|
|
|
| 9919 |
'raw_llm_output', 'raw_output', 'raw_llm', 'deterministic', 'llm_output',
|
| 9920 |
'Item 0', 'Item 1', 'Item 2', 'Item 3', 'Item 4', 'Item 5', 'Item 6', 'Item 7', 'Item 8', 'Item 9'
|
| 9921 |
])
|
| 9922 |
+
other_fields = [
|
| 9923 |
+
k for k in analysis_data.keys()
|
| 9924 |
+
if k not in excluded_keys
|
| 9925 |
+
and analysis_data.get(k) is not None
|
| 9926 |
+
and analysis_data.get(k) != ""
|
| 9927 |
+
and not str(k).lower().endswith('_source')
|
| 9928 |
+
]
|
| 9929 |
+
|
| 9930 |
has_key_findings = bool(analysis_data.get('key_findings'))
|
| 9931 |
if has_key_findings or other_fields:
|
| 9932 |
st.markdown(f"### 📊 {t('additional_analysis_results', 'Additional Analysis Results')}")
|
utils/analysis_engine.py
CHANGED
|
@@ -2623,12 +2623,14 @@ class PromptAnalyzer:
|
|
| 2623 |
except Exception:
|
| 2624 |
configured_model = None
|
| 2625 |
|
|
|
|
| 2626 |
preferred_models = []
|
| 2627 |
for m in (
|
| 2628 |
-
configured_model,
|
| 2629 |
"gemini-2.5-pro",
|
| 2630 |
"gemini-2.5-flash",
|
|
|
|
| 2631 |
"gemini-2.0-flash",
|
|
|
|
| 2632 |
"gemini-1.5-pro",
|
| 2633 |
"gemini-1.5-flash",
|
| 2634 |
"gemini-1.5-pro-002",
|
|
@@ -2728,43 +2730,63 @@ class PromptAnalyzer:
|
|
| 2728 |
"""Extract steps dynamically from prompt text"""
|
| 2729 |
try:
|
| 2730 |
steps = []
|
| 2731 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2732 |
# Use regex to find step patterns
|
| 2733 |
step_pattern = r'Step\s+(\d+):\s*([^\n]+)'
|
| 2734 |
matches = re.findall(step_pattern, prompt_text, re.IGNORECASE)
|
| 2735 |
-
|
| 2736 |
-
# Debug: Log what we found
|
| 2737 |
self.logger.info(f"DEBUG: Found {len(matches)} step matches in prompt")
|
| 2738 |
for i, (step_num, step_title) in enumerate(matches):
|
| 2739 |
self.logger.info(f"DEBUG: Match {i+1}: Step {step_num}: {step_title[:100]}...")
|
| 2740 |
-
|
| 2741 |
for step_num, step_title in matches:
|
| 2742 |
-
|
|
|
|
| 2743 |
start_pos = prompt_text.find(f"Step {step_num}:")
|
|
|
|
|
|
|
|
|
|
| 2744 |
if start_pos == -1:
|
| 2745 |
continue
|
| 2746 |
-
|
| 2747 |
-
|
| 2748 |
-
|
| 2749 |
-
|
| 2750 |
-
description = prompt_text[start_pos:].strip()
|
| 2751 |
else:
|
| 2752 |
-
description = prompt_text[start_pos:
|
| 2753 |
-
|
| 2754 |
-
#
|
| 2755 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2756 |
if len(description) > max_description_length:
|
| 2757 |
description = description[:max_description_length] + "... [truncated for processing efficiency]"
|
| 2758 |
-
|
| 2759 |
steps.append({
|
| 2760 |
-
'number':
|
| 2761 |
-
'title':
|
| 2762 |
'description': description
|
| 2763 |
})
|
| 2764 |
-
|
| 2765 |
self.logger.info(f"Extracted {len(steps)} steps from prompt")
|
| 2766 |
return steps
|
| 2767 |
-
|
| 2768 |
except Exception as e:
|
| 2769 |
self.logger.error(f"Error extracting steps from prompt: {str(e)}")
|
| 2770 |
return []
|
|
@@ -8250,24 +8272,107 @@ class AnalysisEngine:
|
|
| 8250 |
except Exception as _e:
|
| 8251 |
self.logger.warning(f"Could not build Step 1 visualizations: {_e}")
|
| 8252 |
|
| 8253 |
-
# Enhanced Step 2 processing with real issue analysis
|
| 8254 |
try:
|
| 8255 |
for i, sr in enumerate(step_results):
|
| 8256 |
if sr and sr.get('step_number') == 2:
|
| 8257 |
-
|
| 8258 |
-
sr['identified_issues'] =
|
| 8259 |
-
sr['soil_issues'] = soil_issues
|
| 8260 |
-
sr['leaf_issues'] = leaf_issues
|
| 8261 |
-
sr['total_issues'] = len(all_issues)
|
| 8262 |
-
sr['issues_source'] = '
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8263 |
step_results[i] = sr
|
| 8264 |
break
|
| 8265 |
except Exception as _e:
|
| 8266 |
self.logger.warning(f"Could not build Step 2 issues: {_e}")
|
| 8267 |
|
| 8268 |
-
# Step
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8269 |
|
| 8270 |
-
# Enhanced Step 5 processing with complete economic forecast
|
| 8271 |
try:
|
| 8272 |
for i, sr in enumerate(step_results):
|
| 8273 |
if sr and sr.get('step_number') == 5:
|
|
@@ -8296,7 +8401,26 @@ class AnalysisEngine:
|
|
| 8296 |
650, 750, scenario_name
|
| 8297 |
)
|
| 8298 |
scenario_data['yearly_data'] = yearly_data
|
| 8299 |
-
sr['economic_forecast_source'] = '
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8300 |
step_results[i] = sr
|
| 8301 |
self.logger.info(f"Injected complete economic forecast with yearly_data into Step 5")
|
| 8302 |
break
|
|
|
|
| 2623 |
except Exception:
|
| 2624 |
configured_model = None
|
| 2625 |
|
| 2626 |
+
# Prefer latest high-quality Gemini models first (probe verifies availability).
|
| 2627 |
preferred_models = []
|
| 2628 |
for m in (
|
|
|
|
| 2629 |
"gemini-2.5-pro",
|
| 2630 |
"gemini-2.5-flash",
|
| 2631 |
+
configured_model,
|
| 2632 |
"gemini-2.0-flash",
|
| 2633 |
+
"gemini-2.0-pro",
|
| 2634 |
"gemini-1.5-pro",
|
| 2635 |
"gemini-1.5-flash",
|
| 2636 |
"gemini-1.5-pro-002",
|
|
|
|
| 2730 |
"""Extract steps dynamically from prompt text"""
|
| 2731 |
try:
|
| 2732 |
steps = []
|
| 2733 |
+
# Known short titles when the prompt dumps the full instruction after "Step N:"
|
| 2734 |
+
default_titles = {
|
| 2735 |
+
1: "Your Soil and Leaf Test Results and Analysis",
|
| 2736 |
+
2: "Diagnose Agronomic Issues",
|
| 2737 |
+
3: "Recommend Solutions",
|
| 2738 |
+
4: "Regenerative Agriculture Integration",
|
| 2739 |
+
5: "Economic Impact Forecast",
|
| 2740 |
+
6: "Yield Forecast Graph",
|
| 2741 |
+
}
|
| 2742 |
+
|
| 2743 |
# Use regex to find step patterns
|
| 2744 |
step_pattern = r'Step\s+(\d+):\s*([^\n]+)'
|
| 2745 |
matches = re.findall(step_pattern, prompt_text, re.IGNORECASE)
|
| 2746 |
+
|
|
|
|
| 2747 |
self.logger.info(f"DEBUG: Found {len(matches)} step matches in prompt")
|
| 2748 |
for i, (step_num, step_title) in enumerate(matches):
|
| 2749 |
self.logger.info(f"DEBUG: Match {i+1}: Step {step_num}: {step_title[:100]}...")
|
| 2750 |
+
|
| 2751 |
for step_num, step_title in matches:
|
| 2752 |
+
num = int(step_num)
|
| 2753 |
+
# Extract full block for description
|
| 2754 |
start_pos = prompt_text.find(f"Step {step_num}:")
|
| 2755 |
+
if start_pos == -1:
|
| 2756 |
+
start_pos = re.search(rf'Step\s+{step_num}:', prompt_text, re.I)
|
| 2757 |
+
start_pos = start_pos.start() if start_pos else -1
|
| 2758 |
if start_pos == -1:
|
| 2759 |
continue
|
| 2760 |
+
|
| 2761 |
+
next_m = re.search(rf'Step\s+{num + 1}:', prompt_text[start_pos + 1:], re.I)
|
| 2762 |
+
if next_m:
|
| 2763 |
+
description = prompt_text[start_pos:start_pos + 1 + next_m.start()].strip()
|
|
|
|
| 2764 |
else:
|
| 2765 |
+
description = prompt_text[start_pos:].strip()
|
| 2766 |
+
|
| 2767 |
+
# Short UI title: first clause only (avoid dumping entire prompt into progress/UI)
|
| 2768 |
+
raw_title = step_title.strip()
|
| 2769 |
+
short_title = raw_title
|
| 2770 |
+
for sep in (". ", ": ", " — ", " - "):
|
| 2771 |
+
if sep in short_title and short_title.index(sep) < 90:
|
| 2772 |
+
short_title = short_title.split(sep, 1)[0].strip()
|
| 2773 |
+
break
|
| 2774 |
+
if len(short_title) > 90 or len(raw_title) > 120:
|
| 2775 |
+
short_title = default_titles.get(num, short_title[:80].rstrip() + "…")
|
| 2776 |
+
|
| 2777 |
+
max_description_length = 2500
|
| 2778 |
if len(description) > max_description_length:
|
| 2779 |
description = description[:max_description_length] + "... [truncated for processing efficiency]"
|
| 2780 |
+
|
| 2781 |
steps.append({
|
| 2782 |
+
'number': num,
|
| 2783 |
+
'title': short_title,
|
| 2784 |
'description': description
|
| 2785 |
})
|
| 2786 |
+
|
| 2787 |
self.logger.info(f"Extracted {len(steps)} steps from prompt")
|
| 2788 |
return steps
|
| 2789 |
+
|
| 2790 |
except Exception as e:
|
| 2791 |
self.logger.error(f"Error extracting steps from prompt: {str(e)}")
|
| 2792 |
return []
|
|
|
|
| 8272 |
except Exception as _e:
|
| 8273 |
self.logger.warning(f"Could not build Step 1 visualizations: {_e}")
|
| 8274 |
|
| 8275 |
+
# Enhanced Step 2 processing with real issue analysis + readable narrative
|
| 8276 |
try:
|
| 8277 |
for i, sr in enumerate(step_results):
|
| 8278 |
if sr and sr.get('step_number') == 2:
|
| 8279 |
+
issues = self._build_step2_issues(soil_params, leaf_params, all_issues)
|
| 8280 |
+
sr['identified_issues'] = issues
|
| 8281 |
+
sr['soil_issues'] = soil_issues or []
|
| 8282 |
+
sr['leaf_issues'] = leaf_issues or []
|
| 8283 |
+
sr['total_issues'] = len(all_issues) if all_issues else max(0, len(issues) - 1)
|
| 8284 |
+
sr['issues_source'] = 'standards_comparison'
|
| 8285 |
+
|
| 8286 |
+
# Replace empty/placeholder LLM text with real diagnosis narrative
|
| 8287 |
+
empty_summaries = {
|
| 8288 |
+
'', 'analysis completed', 'detailed analysis not available',
|
| 8289 |
+
'analysis completed using fallback processing',
|
| 8290 |
+
}
|
| 8291 |
+
sum_l = str(sr.get('summary') or '').strip().lower()
|
| 8292 |
+
det_l = str(sr.get('detailed_analysis') or '').strip().lower()
|
| 8293 |
+
real_issues = [x for x in issues if isinstance(x, dict) and x.get('type') != 'Summary']
|
| 8294 |
+
soil_n = len([x for x in real_issues if x.get('type') == 'Soil'])
|
| 8295 |
+
leaf_n = len([x for x in real_issues if x.get('type') == 'Leaf'])
|
| 8296 |
+
|
| 8297 |
+
if sum_l in empty_summaries or not sum_l:
|
| 8298 |
+
sr['summary'] = (
|
| 8299 |
+
f"Issue diagnosis identified {len(real_issues)} agronomic concern(s) "
|
| 8300 |
+
f"({soil_n} soil, {leaf_n} leaf) against Malaysian oil palm standards."
|
| 8301 |
+
)
|
| 8302 |
+
if det_l in empty_summaries or not det_l or len(det_l) < 40:
|
| 8303 |
+
lines = [
|
| 8304 |
+
"The following diagnosis is based on measured soil and leaf averages "
|
| 8305 |
+
"compared with Malaysian (MPOB-aligned) adequacy ranges.",
|
| 8306 |
+
"",
|
| 8307 |
+
]
|
| 8308 |
+
for iss in real_issues[:12]:
|
| 8309 |
+
param = iss.get('parameter', 'Parameter')
|
| 8310 |
+
sev = iss.get('severity', 'Unknown')
|
| 8311 |
+
val = iss.get('current_value', 'N/A')
|
| 8312 |
+
cause = iss.get('cause', '')
|
| 8313 |
+
impact = iss.get('impact', '')
|
| 8314 |
+
rec = iss.get('recommendation', '')
|
| 8315 |
+
lines.append(f"**{param}** ({iss.get('type', 'Issue')}) — {sev}")
|
| 8316 |
+
lines.append(f"- Current average: {val}")
|
| 8317 |
+
if cause:
|
| 8318 |
+
lines.append(f"- Cause: {cause}")
|
| 8319 |
+
if impact:
|
| 8320 |
+
lines.append(f"- Impact: {impact}")
|
| 8321 |
+
if rec:
|
| 8322 |
+
lines.append(f"- Action: {rec}")
|
| 8323 |
+
lines.append("")
|
| 8324 |
+
if not real_issues:
|
| 8325 |
+
lines.append(
|
| 8326 |
+
"No critical out-of-range parameters were flagged from the current averages. "
|
| 8327 |
+
"Continue routine monitoring and maintain balanced fertilization."
|
| 8328 |
+
)
|
| 8329 |
+
sr['detailed_analysis'] = "\n".join(lines).strip()
|
| 8330 |
+
if not sr.get('key_findings'):
|
| 8331 |
+
sr['key_findings'] = [
|
| 8332 |
+
f"{iss.get('parameter')}: {iss.get('severity')}"
|
| 8333 |
+
for iss in real_issues[:6]
|
| 8334 |
+
] or ["No critical issues flagged from standards comparison"]
|
| 8335 |
step_results[i] = sr
|
| 8336 |
break
|
| 8337 |
except Exception as _e:
|
| 8338 |
self.logger.warning(f"Could not build Step 2 issues: {_e}")
|
| 8339 |
|
| 8340 |
+
# Step 3: enrich empty recommendations from identified issues / recommendations list
|
| 8341 |
+
try:
|
| 8342 |
+
for i, sr in enumerate(step_results):
|
| 8343 |
+
if sr and sr.get('step_number') == 3:
|
| 8344 |
+
empty = {
|
| 8345 |
+
'', 'analysis completed', 'detailed analysis not available',
|
| 8346 |
+
'solution recommendations prepared',
|
| 8347 |
+
}
|
| 8348 |
+
sum_l = str(sr.get('summary') or '').strip().lower()
|
| 8349 |
+
det_l = str(sr.get('detailed_analysis') or '').strip().lower()
|
| 8350 |
+
sols = sr.get('solution_options') or sr.get('specific_recommendations') or []
|
| 8351 |
+
if (sum_l in empty or det_l in empty or not sols) and recommendations:
|
| 8352 |
+
if sum_l in empty:
|
| 8353 |
+
sr['summary'] = (
|
| 8354 |
+
f"Solution framework prepared with {len(recommendations)} recommendation area(s) "
|
| 8355 |
+
f"prioritized from diagnosed soil and leaf gaps."
|
| 8356 |
+
)
|
| 8357 |
+
if det_l in empty or len(det_l) < 40:
|
| 8358 |
+
lines = ["Recommended corrective actions based on diagnosed nutrient and soil issues:", ""]
|
| 8359 |
+
for idx, rec in enumerate(recommendations[:10], 1):
|
| 8360 |
+
if isinstance(rec, dict):
|
| 8361 |
+
action = rec.get('action') or rec.get('recommendation') or rec.get('title') or str(rec)
|
| 8362 |
+
timing = rec.get('timeline') or rec.get('priority') or ''
|
| 8363 |
+
lines.append(f"{idx}. {action}" + (f" ({timing})" if timing else ""))
|
| 8364 |
+
else:
|
| 8365 |
+
lines.append(f"{idx}. {rec}")
|
| 8366 |
+
sr['detailed_analysis'] = "\n".join(lines)
|
| 8367 |
+
if not sols:
|
| 8368 |
+
sr['solution_options'] = recommendations
|
| 8369 |
+
sr['specific_recommendations'] = recommendations
|
| 8370 |
+
step_results[i] = sr
|
| 8371 |
+
break
|
| 8372 |
+
except Exception as _e:
|
| 8373 |
+
self.logger.warning(f"Could not enrich Step 3 solutions: {_e}")
|
| 8374 |
|
| 8375 |
+
# Enhanced Step 5 processing with complete economic forecast + narrative
|
| 8376 |
try:
|
| 8377 |
for i, sr in enumerate(step_results):
|
| 8378 |
if sr and sr.get('step_number') == 5:
|
|
|
|
| 8401 |
650, 750, scenario_name
|
| 8402 |
)
|
| 8403 |
scenario_data['yearly_data'] = yearly_data
|
| 8404 |
+
sr['economic_forecast_source'] = 'calculated'
|
| 8405 |
+
empty = {'', 'analysis completed', 'detailed analysis not available', 'economic impact assessment prepared'}
|
| 8406 |
+
sum_l = str(sr.get('summary') or '').strip().lower()
|
| 8407 |
+
det_l = str(sr.get('detailed_analysis') or '').strip().lower()
|
| 8408 |
+
if economic_forecast and (sum_l in empty or det_l in empty):
|
| 8409 |
+
land = economic_forecast.get('land_size_hectares') or land_yield_data.get('land_size', 'N/A')
|
| 8410 |
+
yld = economic_forecast.get('current_yield_tonnes_per_ha') or land_yield_data.get('current_yield', 'N/A')
|
| 8411 |
+
scenarios = list((economic_forecast.get('scenarios') or {}).keys())
|
| 8412 |
+
if sum_l in empty:
|
| 8413 |
+
sr['summary'] = (
|
| 8414 |
+
f"Economic impact forecast prepared for {land} ha at a current yield baseline "
|
| 8415 |
+
f"of {yld} t/ha, with {len(scenarios)} investment scenario(s)."
|
| 8416 |
+
)
|
| 8417 |
+
if det_l in empty:
|
| 8418 |
+
sr['detailed_analysis'] = (
|
| 8419 |
+
"Five-year economic projections compare low, medium, and high investment "
|
| 8420 |
+
"pathways for nutrient correction. Tables below show yearly yield, cost, "
|
| 8421 |
+
"revenue, and net position for each scenario. Use these figures to prioritize "
|
| 8422 |
+
"affordable interventions that close the largest nutrient gaps first."
|
| 8423 |
+
)
|
| 8424 |
step_results[i] = sr
|
| 8425 |
self.logger.info(f"Injected complete economic forecast with yearly_data into Step 5")
|
| 8426 |
break
|
utils/streamlit_runtime.py
CHANGED
|
@@ -48,12 +48,30 @@ class AnalysisProgressReporter:
|
|
| 48 |
working=message,
|
| 49 |
)
|
| 50 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 51 |
def on_step_start(self, step_number: int, step_title: str, total_steps: int) -> None:
|
| 52 |
self._total_steps = max(total_steps, 1)
|
| 53 |
fraction = max(0.0, (step_number - 1) / self._total_steps)
|
| 54 |
-
self.
|
| 55 |
-
|
| 56 |
-
|
|
|
|
|
|
|
|
|
|
| 57 |
)
|
| 58 |
|
| 59 |
def on_stream_chunk(self, step_number: int, chars_received: int) -> None:
|
|
@@ -63,7 +81,7 @@ class AnalysisProgressReporter:
|
|
| 63 |
self._last_flush = now
|
| 64 |
fraction = min(0.99, ((step_number - 1) + min(chars_received / 4000.0, 0.95)) / self._total_steps)
|
| 65 |
progress = int(self.base_progress + (self.progress_span * fraction))
|
| 66 |
-
stream_msg = f"Receiving AI output
|
| 67 |
self._update(
|
| 68 |
progress=progress,
|
| 69 |
status=f"🤖 **AI Step {step_number}/{self._total_steps}:** {stream_msg}",
|
|
@@ -73,9 +91,10 @@ class AnalysisProgressReporter:
|
|
| 73 |
|
| 74 |
def on_step_complete(self, step_number: int) -> None:
|
| 75 |
fraction = min(1.0, step_number / self._total_steps)
|
| 76 |
-
self.
|
| 77 |
-
|
| 78 |
-
|
|
|
|
| 79 |
)
|
| 80 |
|
| 81 |
def _update(
|
|
@@ -90,13 +109,17 @@ class AnalysisProgressReporter:
|
|
| 90 |
self.progress_bar.progress(min(max(int(progress), 0), 100))
|
| 91 |
if self.status_text is not None:
|
| 92 |
self.status_text.markdown(status)
|
|
|
|
| 93 |
if self.working_indicator is not None and working:
|
| 94 |
-
|
|
|
|
|
|
|
|
|
|
| 95 |
if self.stream_indicator is not None and stream:
|
| 96 |
self.stream_indicator.caption(stream)
|
| 97 |
if self.time_estimate is not None:
|
| 98 |
self.time_estimate.markdown(
|
| 99 |
-
"⏱️
|
| 100 |
)
|
| 101 |
if self.step_indicator is not None:
|
| 102 |
self.step_indicator.markdown("📋 AI analysis in progress")
|
|
|
|
| 48 |
working=message,
|
| 49 |
)
|
| 50 |
|
| 51 |
+
@staticmethod
|
| 52 |
+
def _short_title(step_title: str, max_len: int = 72) -> str:
|
| 53 |
+
"""Collapse long step prompts into a short progress label (no duplicate walls of text)."""
|
| 54 |
+
text = " ".join(str(step_title or "").split())
|
| 55 |
+
if not text:
|
| 56 |
+
return "Analysis"
|
| 57 |
+
# Prefer text before first period / colon after a short phrase
|
| 58 |
+
for sep in (". ", ": ", " — ", " - "):
|
| 59 |
+
if sep in text and text.index(sep) < max_len:
|
| 60 |
+
text = text.split(sep, 1)[0].strip()
|
| 61 |
+
break
|
| 62 |
+
if len(text) > max_len:
|
| 63 |
+
text = text[: max_len - 1].rstrip() + "…"
|
| 64 |
+
return text
|
| 65 |
+
|
| 66 |
def on_step_start(self, step_number: int, step_title: str, total_steps: int) -> None:
|
| 67 |
self._total_steps = max(total_steps, 1)
|
| 68 |
fraction = max(0.0, (step_number - 1) / self._total_steps)
|
| 69 |
+
short = self._short_title(step_title)
|
| 70 |
+
# Single concise status line — do NOT repeat the full prompt in Processing
|
| 71 |
+
self._update(
|
| 72 |
+
progress=self.base_progress + (self.progress_span * fraction),
|
| 73 |
+
status=f"🤖 **AI Step {step_number}/{total_steps}:** {short}",
|
| 74 |
+
working=f"Step {step_number} of {total_steps} — generating analysis…",
|
| 75 |
)
|
| 76 |
|
| 77 |
def on_stream_chunk(self, step_number: int, chars_received: int) -> None:
|
|
|
|
| 81 |
self._last_flush = now
|
| 82 |
fraction = min(0.99, ((step_number - 1) + min(chars_received / 4000.0, 0.95)) / self._total_steps)
|
| 83 |
progress = int(self.base_progress + (self.progress_span * fraction))
|
| 84 |
+
stream_msg = f"Receiving AI output… {chars_received:,} characters"
|
| 85 |
self._update(
|
| 86 |
progress=progress,
|
| 87 |
status=f"🤖 **AI Step {step_number}/{self._total_steps}:** {stream_msg}",
|
|
|
|
| 91 |
|
| 92 |
def on_step_complete(self, step_number: int) -> None:
|
| 93 |
fraction = min(1.0, step_number / self._total_steps)
|
| 94 |
+
self._update(
|
| 95 |
+
progress=self.base_progress + (self.progress_span * fraction),
|
| 96 |
+
status=f"✅ **AI Step {step_number}/{self._total_steps}** complete",
|
| 97 |
+
working=f"Step {step_number} finished",
|
| 98 |
)
|
| 99 |
|
| 100 |
def _update(
|
|
|
|
| 109 |
self.progress_bar.progress(min(max(int(progress), 0), 100))
|
| 110 |
if self.status_text is not None:
|
| 111 |
self.status_text.markdown(status)
|
| 112 |
+
# Avoid duplicating the same long status under Processing
|
| 113 |
if self.working_indicator is not None and working:
|
| 114 |
+
if working.strip() != status.replace("**", "").strip():
|
| 115 |
+
self.working_indicator.markdown(f"🔄 **Processing:** {working}")
|
| 116 |
+
else:
|
| 117 |
+
self.working_indicator.markdown("🔄 **Processing:** analysis in progress…")
|
| 118 |
if self.stream_indicator is not None and stream:
|
| 119 |
self.stream_indicator.caption(stream)
|
| 120 |
if self.time_estimate is not None:
|
| 121 |
self.time_estimate.markdown(
|
| 122 |
+
"⏱️ **Keep this page open** during analysis (typically 5–8 minutes)."
|
| 123 |
)
|
| 124 |
if self.step_indicator is not None:
|
| 125 |
self.step_indicator.markdown("📋 AI analysis in progress")
|