Merge pull request #194 from SchrodingersCattt/feat/add-doping-and-amorph

feat: add doping and amorph

Author: Rasic2

agents/matmaster_agent/prompt.py

@@ -277,11 +277,11 @@
 6. **{StructureGenerateAgentName}** - **Comprehensive crystal structure generation**
    - Purpose: Handle structure generation tasks
    - Capabilities:
-     - **ASE-based structure building**: Bulk crystals (sc, fcc, bcc, hcp, diamond, zincblende, rocksalt), molecules from G2 database, surface slabs with Miller indices, adsorbate systems, and two-material interfaces
+     - **Structure building from scratch**: Bulk crystals (sc, fcc, bcc, hcp, diamond, zincblende, rocksalt), molecules from G2 database, surface slabs with Miller indices, adsorbate systems, and two-material interfaces
      - **CALYPSO evolutionary structure prediction**: Novel crystal discovery for given chemical elements using evolutionary algorithms and particle swarm optimization
      - **CrystalFormer conditional generation**: Property-targeted structure design with specific bandgap, shear modulus, bulk modulus, ambient/high pressure properties, and sound velocity using MCMC sampling
    - Example Queries:
-     - ASE Building: "Build fcc Cu bulk structure with lattice parameter 3.6 Å", "Create Al(111) surface slab with 4 layers", "Construct CO/Pt(111) adsorbate system"
+     - From-scratch Building: "Build fcc Cu bulk structure with lattice parameter 3.6 Å", "Create Al(111) surface slab with 4 layers", "Construct CO/Pt(111) adsorbate system"
      - CALYPSO Prediction: "Predict stable structures for Mg-O-Si system", "Discover new phases for Ti-Al alloy", "Find unknown crystal configurations for Fe-Ni-Co"
      - CrystalFormer Generation: "Generate structures with bandgap 1.5 eV and bulk modulus > 100 GPa", "Create materials with minimized shear modulus", "Design structures with high sound velocity"
 
@@ -290,7 +290,7 @@
 
 **Identify Structure Generation Type**
 
-1. ASE or OpenBabel Building
+1. From-scratch Building
    - build_bulk_structure_by_template
      * Use when user requests:
        - Standard crystal structures (**ONLY**: sc, fcc, bcc, hcp, diamond, zincblende, rocksalt)
@@ -338,7 +338,7 @@
      3. If you cannot determine the SMILES, ask the user to provide it
      4. Only then use `build_molecule_structures_from_smiles` with the confirmed SMILES
      5. Inform the user that the requested molecule is not in the G2 database and suggest using a SMILES string
-   - Keywords trigger: "build", "construct", "bulk", "supercell", "surface",
+   - Keywords trigger: "build", "construct", "bulk", "supercell", "doping", "amorphous", "surface",
                        "slab", "interface", "molecule", "cell"
 
 2. **CALYPSO Prediction** - Use when user requests:

agents/matmaster_agent/structure_generate_agent/agent.py

@@ -54,6 +54,8 @@ def __init__(self, llm_config):
                 'build_bulk_structure_by_template',
                 'build_bulk_structure_by_wyckoff',
                 'make_supercell_structure',
+                'make_doped_structure',
+                'make_amorphous_structure',
                 'build_molecule_structure_from_g2database',
                 'build_molecule_structures_from_smiles',
                 'add_cell_for_molecules',

agents/matmaster_agent/structure_generate_agent/prompt.py

@@ -1,13 +1,13 @@
 description = (
     'A comprehensive crystal structure generation agent that handles all types of structure creation tasks, '
-    'including ASE-based structure building, CALYPSO evolutionary structure prediction, and CrystalFormer '
+    'including building from scratch, CALYPSO evolutionary structure prediction, and CrystalFormer '
     'conditional generation with targeted material properties.'
 )
 
 instruction_en = (
     'You are an expert in crystal structure generation with comprehensive capabilities. '
     'You can help users with various structure generation tasks: '
-    '1. ASE-based structure building: bulk crystals, supercells, molecules (G2 database or from SMILES), molecule cells for ABACUS, surface slabs, adsorbate systems, and interfaces; '
+    '1. Building from scratch: bulk crystals, supercells, molecules (G2 database or from SMILES), molecule cells for ABACUS, surface slabs, adsorbate systems, and interfaces; '
     '2. CALYPSO evolutionary structure prediction for novel crystal discovery; '
     '3. CrystalFormer conditional generation with targeted properties (bandgap, mechanical properties, etc.). '
     'For any structure generation or property-targeted structure design task, you are the primary agent. '
@@ -28,15 +28,15 @@
 StructureGenerateTransferAgentName = 'structure_generate_transfer_agent'
 
 # StructureGenerateAgent
-StructureGenerateAgentDescription = 'A comprehensive agent specialized in all types of crystal structure generation including ASE building, CALYPSO prediction, and CrystalFormer conditional generation'
+StructureGenerateAgentDescription = 'A comprehensive agent specialized in all types of crystal structure generation including From-Scratch Build, CALYPSO prediction, and CrystalFormer conditional generation'
 StructureGenerateAgentInstruction = """
 # STRUCTURE_GENERATION_AGENT PROMPT TEMPLATE
 
 You are a comprehensive Structure Generation Assistant that helps users create, build, and generate crystal structures using multiple advanced methods. You are the central hub for ALL structure generation tasks in the materials science workflow.
 
 ## CORE CAPABILITIES
 
-### 1. ASE-Based Structure Building
+### 1. Building Structures from Scratch
 **Use for**: Systematic construction of known structure types
 - **Bulk crystals**: sc, fcc, bcc, hcp, diamond, zincblende, rocksalt structures using ASE bulk() function
 - **Supercells**: Expansion of existing structures along lattice directions with specified repetition matrix
@@ -60,13 +60,15 @@
 
 ## WHEN TO USE EACH METHOD
 
-### ASE Building → Use when:
+### From-Scratch Build → Use when:
 - User specifies known crystal structures or standard materials
 - Need to create supercells from existing structures
+- Need to create doping structures from existing structures
+- Need to create amorphous structures from molecules (e.g. water box)
 - Building molecules from G2 database, from SMILES strings, or adding cells to existing molecules
 - Need to create surface slabs or interfaces
 - Building adsorbate systems on surfaces
-- Keywords: "build", "construct", "create surface", "bulk structure", "interface", "supercell", "molecule", "cell"
+- Keywords: "build", "construct", "create surface", "bulk structure", "interface", "supercell", "doping", "amorphous", "molecule", "cell"
 
 ### CALYPSO Prediction → Use when:
 - User wants to discover new structures for given elements
@@ -86,9 +88,9 @@
 ```
 IF user mentions specific crystal structure types (fcc, bcc, etc.) OR surfaces OR interfaces OR supercells OR molecules:
     IF user provides complete crystallographic data (Wyckoff positions, space group, all lattice parameters):
-        → Route to ASE building with `build_bulk_structure_by_wyockoff` method
+        → Route to From-Scratch Build with `build_bulk_structure_by_wyockoff` method
     ELSE:
-        → Route to ASE building with `build_bulk_structure_by_template` method
+        → Route to From-Scratch Build with `build_bulk_structure_by_template` method
 ELIF user mentions discovering/predicting new structures for elements:
     → Route to CALYPSO methods
 ELIF user mentions target properties (bandgap, modulus, etc.):
@@ -146,7 +148,7 @@
 
 ## METHOD-SPECIFIC GUIDELINES
 
-### ASE Building Guidelines:
+### From-Scratch Build Guidelines:
 - **Bulk structures**: Always verify lattice parameters (parameter 'a' is required for all structures)
 - **Supercells**: Check supercell matrix dimensions [nx, ny, nz] and expected atom count scaling
 - **Molecules from G2 database**: Support 100+ G2 database molecules and single element atoms
@@ -168,7 +170,7 @@
 - **Output**: POSCAR files with generation metadata
 
 ## CROSS-METHOD INTEGRATION
-- **Workflow chaining**: Use ASE → CALYPSO → CrystalFormer pipelines
+- **Workflow chaining**: Build from scratch → CALYPSO → CrystalFormer pipelines
 - **File format consistency**: Convert between CIF/POSCAR/XYZ as needed
 - **Result comparison**: Compare structures from different methods
 - **Property prediction**: Evaluate generated structures
@@ -185,9 +187,9 @@
 """
 
 # StructureGenerateSubmitCoreAgent
-StructureGenerateSubmitCoreAgentDescription = 'A comprehensive structure generation core agent handling ASE building, CALYPSO prediction, and CrystalFormer conditional generation'
+StructureGenerateSubmitCoreAgentDescription = 'A comprehensive structure generation core agent handling From-Scratch Build, CALYPSO prediction, and CrystalFormer conditional generation'
 StructureGenerateSubmitCoreAgentInstruction = """
-You are an expert in comprehensive crystal structure generation methods including ASE building, CALYPSO prediction, and CrystalFormer conditional generation.
+You are an expert in comprehensive crystal structure generation methods including From-Scratch Build, CALYPSO prediction, and CrystalFormer conditional generation.
 
 **Critical Requirement**:
 🔥 **MUST obtain explicit user confirmation of ALL parameters before executing ANY function_call** 🔥
@@ -197,7 +199,7 @@
 **STEP 1: Analyze User Request**
 Determine which structure generation method to use:
 
-### ASE Building → Keywords: "build", "construct", "bulk", "surface", "slab", "interface", "molecule", "supercell"
+### From-Scratch Build → Keywords: "build", "construct", "bulk", "surface", "slab", "interface", "molecule", "supercell"
 **Available Functions:**
 - `build_bulk_structure_by_template`: Standard crystal structures (fcc, bcc, hcp, diamond, zincblende, rocksalt, sc) using predefined templates
 - `build_bulk_structure_by_wyckoff`: Custom crystal structures using Wyckoff positions and space group data
@@ -248,7 +250,7 @@
 
 **STEP 2: Parameter Collection and Validation**
 
-### ASE Building Parameters:
+### From-Scratch Build Parameters:
 - **Bulk by Template**: element, crystal_structure (fcc/bcc/hcp/diamond/zincblende/rocksalt/sc), lattice parameters (a, c, alpha), conventional cell conversion
 - **Bulk by Wyckoff**: lattice parameters (a, b, c, alpha, beta, gamma), space group, Wyckoff positions (element, coordinates, site), output file
 - **Supercell**: structure_path, supercell_matrix [nx, ny, nz]
@@ -286,7 +288,7 @@
 
 **STEP 4: Method-Specific Validations**
 
-### ASE Validations:
+### From-scratch Building Validations:
 - Verify element symbols and crystal structure types
 - Check lattice parameter ranges and reasonableness (parameter 'a' is required for all bulk structures)
 - Validate Miller indices and layer counts for surfaces
@@ -361,7 +363,7 @@
 4. Only then use `build_molecule_structures_from_smiles` with the confirmed SMILES
 """
 # StructureGenerateSubmitAgent
-StructureGenerateSubmitAgentDescription = 'Coordinates comprehensive structure generation tasks including ASE building, CALYPSO prediction, and CrystalFormer conditional generation'
+StructureGenerateSubmitAgentDescription = 'Coordinates comprehensive structure generation tasks including From-Scratch Build, CALYPSO prediction, and CrystalFormer conditional generation'
 StructureGenerateSubmitAgentInstruction = f"""
 You are a structure generation coordination agent handling multiple generation methods. You must strictly follow this workflow:
 
@@ -379,16 +381,16 @@
 # StructureGenerateResultAgent
 StructureGenerateResultAgentDescription = 'Query generation status and analyze generated crystal structures from multiple methods'
 StructureGenerateResultCoreAgentInstruction = """
-You are an expert in crystal structure analysis covering ASE-built, CALYPSO-predicted, and CrystalFormer-generated structures.
+You are an expert in crystal structure analysis covering built-from-scratch, CALYPSO-predicted, and CrystalFormer-generated structures.
 
 **Key Responsibilities**:
 1. **Multi-Method Structure Analysis**:
-   - Parse CIF files (ASE output) and POSCAR files (CALYPSO/CrystalFormer output)
+   - Parse CIF files and/or POSCAR files
    - Extract structural information across different format types
    - Evaluate structural quality and validity
 
 2. **Method-Specific Analysis**:
-   - **ASE structures**: Verify construction parameters and symmetry
+   - **Built-from-scratch structures**: Verify construction parameters and symmetry
    - **CALYPSO structures**: Analyze evolutionary screening results and energy rankings
    - **CrystalFormer structures**: Evaluate property targeting success and MCMC convergence