memory/generate_examples.sh at main · FastBuilderAI/memory · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
#!/bin/bash
set -e

cd /Users/pkpro/upperspace3/fastmemory
mkdir -p example/health_science example/robotics example/driverless_cars example/audit example/email_analysis example/business_analytics example/world_events

function write_html {
    local folder=$1
    cat << 'EOF' > "$folder/index.html"
<!DOCTYPE html>
<html>
<head>
    <title>FastMemory Visualization</title>
    <script src="https://d3js.org/d3.v7.min.js"></script>
    <style>
        body { font-family: -apple-system, sans-serif; margin: 0; padding: 20px; background: #0f172a; color: white; }
        .node { stroke: #fff; stroke-width: 1.5px; }
        .link { stroke: #334155; stroke-opacity: 0.6; }
        svg { background: #1e293b; border-radius: 8px; width: 100%; height: 800px; }
        .label { font-size: 12px; fill: white; pointer-events: none; }
    </style>
</head>
<body>
    <h2>Clustered Text Memory Graph (Louvain)</h2>
    <div id="chart"></div>
    <script>
        fetch('output.json').then(r => r.json()).then(data => {
            const nodes = [];
            const links = [];

            // Flatten blocks into nodes
            data.forEach((block, i) => {
                nodes.push({ id: block.id, group: i, name: block.name, size: block.node_count * 10 + 10 });
                block.nodes.forEach(n => {
                    nodes.push({ id: n, group: i, name: n, size: 5 });
                    links.push({ source: n, target: block.id, value: 1 });
                });
            });

            const width = document.getElementById('chart').clientWidth;
            const height = 800;

            const simulation = d3.forceSimulation(nodes)
                .force("link", d3.forceLink(links).id(d => d.id).distance(50))
                .force("charge", d3.forceManyBody().strength(-200))
                .force("center", d3.forceCenter(width / 2, height / 2));

            const svg = d3.select("#chart").append("svg")
                .attr("viewBox", [0, 0, width, height]);

            const link = svg.append("g")
                .attr("class", "link")
                .selectAll("line")
                .data(links)
                .join("line");

            const node = svg.append("g")
                .attr("class", "node")
                .selectAll("circle")
                .data(nodes)
                .join("circle")
                .attr("r", d => d.size)
                .attr("fill", d => d3.schemeCategory10[d.group % 10])
                .call(drag(simulation));

            node.append("title").text(d => d.name);

            const label = svg.append("g")
                .attr("class", "label")
                .selectAll("text")
                .data(nodes.filter(n => n.size > 5)) // Only label blocks
                .join("text")
                .text(d => d.name)
                .attr("x", 8)
                .attr("y", 3);

            simulation.on("tick", () => {
                link
                    .attr("x1", d => d.source.x)
                    .attr("y1", d => d.source.y)
                    .attr("x2", d => d.target.x)
                    .attr("y2", d => d.target.y);
                node
                    .attr("cx", d => d.x)
                    .attr("cy", d => d.y);
                label
                    .attr("x", d => d.x + 12)
                    .attr("y", d => d.y + 4);
            });

            function drag(simulation) {
                function dragstarted(event) {
                    if (!event.active) simulation.alphaTarget(0.3).restart();
                    event.subject.fx = event.subject.x;
                    event.subject.fy = event.subject.y;
                }
                function dragged(event) {
                    event.subject.fx = event.x;
                    event.subject.fy = event.y;
                }
                function dragended(event) {
                    if (!event.active) simulation.alphaTarget(0);
                    event.subject.fx = null;
                    event.subject.fy = null;
                }
                return d3.drag()
                    .on("start", dragstarted)
                    .on("drag", dragged)
                    .on("end", dragended);
            }
        });
    </script>
</body>
</html>
EOF
}

# HEALTH SCIENCE (e.g. PubMed QA graph vs Vector)
cat << 'EOF' > example/health_science/input.md
# Health Science: Precision Medicine & Oncology
Original Vector Approach: Embedded clinical trials independently. RAG failed on multi-drug interactions.
FastMemory Approach: Converts Drug mechanisms and patient profiles into ATFs with contextual graph links.

## [ID: ATF_DRUG_001]
**Action:** Administrate_Paclitaxel
**Input:** {Patient_Age, Tumor_Stage}
**Logic:** Inhibit microtubule breakdown.
**Context_Links:** [ATF_SIDE_001], [ATF_MOA_001]

## [ID: ATF_SIDE_001]
**Action:** Monitor_Neuropathy
**Input:** {Patient_Symptoms}
**Logic:** If tingling, reduce dose.
**Context_Links:** [ATF_DRUG_001]

## [ID: ATF_MOA_001]
**Action:** Inhibit_Mitosis
**Input:** {Cellular_Process}
**Logic:** Stops tumor cell division.
**Context_Links:** [ATF_DRUG_001]
EOF

# ROBOTICS
cat << 'EOF' > example/robotics/input.md
# Robotics: Swarm Navigation
Original Vector Approach: Nav-logs embedded. Failed to map conditional failure states (e.g. "If sensor dead, follow neighbour").
FastMemory Approach: Conditionals are parsed as functional ATFs with explicit dependency edges.

## [ID: ATF_NAV_001]
**Action:** Sensor_Fusion_Logic
**Input:** {LidarData, CameraData}
**Logic:** Merge into local costmap.
**Context_Links:** [ATF_FAIL_001]

## [ID: ATF_FAIL_001]
**Action:** Fallback_To_Swarm
**Input:** {Sensor_Status}
**Logic:** If Lidar dead, tracking neighbor bot ID.
**Context_Links:** [ATF_NAV_001]
EOF

# DRIVERLESS CARS
cat << 'EOF' > example/driverless_cars/input.md
# Driverless Cars: Perception & Scenario Trees
Original Vector Approach: Scene graphs clustered via point clouds.
FastMemory Approach: Semantic rules of the road converted to interconnected ATFs.

## [ID: ATF_SCENE_PEDESTRIAN]
**Action:** Yield_To_Pedestrian
**Input:** {Crosswalk_Zone, Velocity}
**Logic:** Must stop if pedestrian in zone.
**Context_Links:** [ATF_ACT_BRAKE]

## [ID: ATF_ACT_BRAKE]
**Action:** Apply_Braking_Force
**Input:** {Deceleration_Target}
**Logic:** Ramp up brake pressure smoothly.
**Context_Links:** [ATF_SCENE_PEDESTRIAN]
EOF

# AUDIT
cat << 'EOF' > example/audit/input.md
# Audit: Financial Fraud Detection (FinQA)
Previous: Flattened ledger parsing with RAG finding textual matches.
FastMemory: ATFs tracing the flow of funds and compliance rules.

## [ID: ATF_COMPLY_SOX]
**Action:** Verify_SOX_Compliance
**Input:** {Report_Q3}
**Logic:** Check signatures from CEO/CFO.
**Context_Links:** [ATF_LEDGER_REVIEW]

## [ID: ATF_LEDGER_REVIEW]
**Action:** Audit_Expense_Anomalies
**Input:** {T_E_Ledger}
**Logic:** Flag if expense > $10k without receipt.
**Context_Links:** [ATF_COMPLY_SOX]
EOF

# EMAIL ANALYSIS
cat << 'EOF' > example/email_analysis/input.md
# Email: Enron Corporate Communications
Previous Approach: NLP classification (BERT).
FastMemory: Email threads modeled as conversational logic blocks.

## [ID: ATF_THREAD_102]
**Action:** Discuss_Offshore_Accounts
**Input:** {Financial_Strategy}
**Logic:** Move assets to subsidiary.
**Context_Links:** [ATF_THREAD_103]

## [ID: ATF_THREAD_103]
**Action:** Authorize_Transfer
**Input:** {Approval_Chain}
**Logic:** CFO signature required.
**Context_Links:** [ATF_THREAD_102]
EOF

# BUSINESS ANALYTICS
cat << 'EOF' > example/business_analytics/input.md
# Business Analytics: Churn Prediction
Previous: Decision trees on tabular data.
FastMemory: Customer behavior rules synthesized into Actionable Text Functions.

## [ID: ATF_CHURN_DETECT]
**Action:** Detect_Usage_Drop
**Input:** {Weekly_Logins}
**Logic:** If login drop > 50%, flag risk.
**Context_Links:** [ATF_RETENTION_OFFER]

## [ID: ATF_RETENTION_OFFER]
**Action:** Trigger_Discount_Email
**Input:** {Customer_ID}
**Logic:** Send 20% off coupon.
**Context_Links:** [ATF_CHURN_DETECT]
EOF

# WORLD EVENTS
cat << 'EOF' > example/world_events/input.md
# World Events: Geopolitical Supply Chain Mapping
Previous Approach: Knowledge Graphs (Neo4j).
FastMemory: Event nodes clustered dynamically by Louvain communities tracking cascading economic effects.

## [ID: ATF_EVENT_PORT_STRIKE]
**Action:** Delay_Logistics
**Input:** {West_Coast_Ports}
**Logic:** Add 14 days to shipping ETAs.
**Context_Links:** [ATF_IMPACT_AUTO]

## [ID: ATF_IMPACT_AUTO]
**Action:** Adjust_Manufacturing_Forecast
**Input:** {Component_Inventory}
**Logic:** Reduce output by 10%.
**Context_Links:** [ATF_EVENT_PORT_STRIKE]
EOF


for domain in health_science robotics driverless_cars audit email_analysis business_analytics world_events; do
    echo "Processing $domain..."
    cargo run -q -- example/$domain/input.md > example/$domain/output.json
    write_html example/$domain

    cat << EOF > example/$domain/README.md
# $domain - FastMemory Example

## Original Huggingface Scenario
In many Huggingface test datasets involving **$domain**, the original approach relies on standard Vector embeddings (RAG) or static Knowledge Graphs. These approaches often fail with multi-hop reasoning or conditional logic because they retrieve flat snippets based on semantic similarity rather than traversing logical dependencies.

## FastMemory "Clustered Text Memory" Approach
Here we convert the dataset rules into **Atomic Text Functions (ATFs)** and use the strict \`rust-louvain\` graph clustering to map out executable semantic structures.

## Reproducible Command
\`\`\`bash
# From fastmemory root:
cargo run -- example/$domain/input.md > example/$domain/output.json
\`\`\`

## Provable Output
\`\`\`json
$(cat example/$domain/output.json | python3 -m json.tool)
\`\`\`

## Visualization
Open \`index.html\` in a browser to view the Force-Directed graph of the clustered logic components generated by the underlying \`rust-louvain\` engine.
EOF

done