calc—_v2
This commit is contained in:
8
main.sh
8
main.sh
@@ -6,11 +6,11 @@
|
||||
|
||||
# 1. 初始化设置
|
||||
# 修改上一级目录权限
|
||||
chmod -R u+w ../Screen
|
||||
chmod -R u+w ../screen
|
||||
|
||||
# 启用 screen 环境 (Python 3.11)
|
||||
source $(conda info --base)/etc/profile.d/conda.sh
|
||||
conda activate screen
|
||||
conda activate ~/anaconda3/envs/screen
|
||||
|
||||
# 设置当前目录为 PYTHONPATH
|
||||
cd py/
|
||||
@@ -31,7 +31,7 @@ python make_sh.py
|
||||
# 2. 切换环境运行 Zeo++
|
||||
echo "============ Stage 2: Zeo++ Calculations ============"
|
||||
conda deactivate
|
||||
conda activate zeo
|
||||
conda activate ~/anaconda3/envs/zeo
|
||||
|
||||
# 进入数据目录执行所有生成的 shell 脚本
|
||||
cd ../data/after_step1
|
||||
@@ -46,7 +46,7 @@ fi
|
||||
echo "============ Stage 3: Data Extraction & Advanced Screening ============"
|
||||
# 切回 screen 环境
|
||||
conda deactivate
|
||||
conda activate screen
|
||||
conda activate ~/anaconda3/envs/screen
|
||||
cd ../../py
|
||||
|
||||
# 提取日志数据
|
||||
|
||||
@@ -1,12 +1,10 @@
|
||||
import os
|
||||
import pandas as pd
|
||||
import math
|
||||
import shutil
|
||||
|
||||
# ================= 配置区域 =================
|
||||
# 定义各阴离子的筛选阈值
|
||||
# perc: Percolation diameter (对应 Step 2, 大于此值)
|
||||
# min_d: Minimum of d (对应 Step 3, 小于此值)
|
||||
# node: Maximum node length (对应 Step 4, 大于此值)
|
||||
THRESHOLDS = {
|
||||
"O": {"perc": 0.50, "min_d": 3.0, "node": 2.2},
|
||||
"S": {"perc": 0.55, "min_d": 3.0, "node": 2.2},
|
||||
@@ -16,7 +14,7 @@ THRESHOLDS = {
|
||||
|
||||
# 路径配置
|
||||
CSV_ROOT_DIR = "../output" # CSV 所在的根目录
|
||||
DATA_SOURCE_DIR = "../data/after_step1" # 原始 CIF 文件所在的根目录 (用于创建链接源)
|
||||
DATA_SOURCE_DIR = "../data/after_step1" # 原始 CIF 文件所在的根目录
|
||||
TARGET_DIR = "../data/after_screening" # 筛选后放置软链接的目标目录
|
||||
|
||||
|
||||
@@ -26,28 +24,21 @@ def check_requirements(row, anion_type):
|
||||
"""
|
||||
检查单行数据是否符合要求
|
||||
"""
|
||||
# 获取该阴离子类型的阈值配置
|
||||
config = THRESHOLDS.get(anion_type)
|
||||
if not config:
|
||||
print(f"Warning: 未知的阴离子类型 {anion_type}，跳过筛选。")
|
||||
return False
|
||||
|
||||
try:
|
||||
# 获取数值 (处理可能的空值或非数字情况)
|
||||
perc = float(row["Percolation Diameter (A)"])
|
||||
min_d = float(row["Minimum of d"])
|
||||
node = float(row["Maximum Node Length (A)"])
|
||||
|
||||
# 检查是否为 NaN
|
||||
if math.isnan(perc) or math.isnan(min_d) or math.isnan(node):
|
||||
return False
|
||||
|
||||
# --- 筛选逻辑 ---
|
||||
# Step 2: 连通孔径 > 阈值
|
||||
# 筛选逻辑
|
||||
c1 = perc > config["perc"]
|
||||
# Step 3: 最短距离 < 3.0 (所有元素目前都是3.0)
|
||||
c2 = min_d < config["min_d"]
|
||||
# Step 4: 扩大锂离子节点 > 阈值
|
||||
c3 = node > config["node"]
|
||||
|
||||
return c1 and c2 and c3
|
||||
@@ -56,39 +47,56 @@ def check_requirements(row, anion_type):
|
||||
return False
|
||||
|
||||
|
||||
def create_symlink(group_name, anion_name, material_id):
|
||||
def create_result_file(group_name, anion_name, material_id):
|
||||
"""
|
||||
创建软链接
|
||||
源: ../data/after_step1/Group/Anion/ID/ID.cif
|
||||
目: ../data/after_screening/Group/Anion/ID.cif
|
||||
创建结果文件 (这里改为直接复制，软链接在跨文件系统或某些环境下可能不稳定，复制更稳妥)
|
||||
如果确实需要软链接，可以将 shutil.copy 换回 os.symlink
|
||||
"""
|
||||
# 1. 构建源文件路径 (必须使用绝对路径以确保软链接在任何地方都有效)
|
||||
# 注意：根据你修改后的 step1，文件在 ID 文件夹内，如 141/141.cif
|
||||
# 1. 构建源文件路径
|
||||
# 正确路径: ../data/after_step1/Group/MaterialID/MaterialID.cif
|
||||
# 例如: ../data/after_step1/S/195819/195819.cif
|
||||
# 注意：如果原本结构是 S+O/S/ID... 这里会自动适配
|
||||
|
||||
# 这里的路径逻辑要非常小心，取决于 extract_data.py 是怎么生成 CSV 目录结构的
|
||||
# 如果 CSV 在 output/S/S.csv -> 对应源文件在 after_step1/S/...
|
||||
# 如果 CSV 在 output/S+O/S/S.csv -> 对应源文件在 after_step1/S+O/S/...
|
||||
|
||||
if group_name == anion_name:
|
||||
# 单阴离子情况 (如 output/S/S.csv -> after_step1/S/ID/ID.cif)
|
||||
rel_source_path = os.path.join(DATA_SOURCE_DIR, group_name, material_id, f"{material_id}.cif")
|
||||
else:
|
||||
# 混合阴离子情况 (如 output/S+O/S/S.csv -> after_step1/S+O/S/ID/ID.cif)
|
||||
rel_source_path = os.path.join(DATA_SOURCE_DIR, group_name, anion_name, material_id, f"{material_id}.cif")
|
||||
|
||||
abs_source_path = os.path.abspath(rel_source_path)
|
||||
|
||||
if not os.path.exists(abs_source_path):
|
||||
print(f"源文件不存在: {abs_source_path}")
|
||||
return
|
||||
|
||||
# 2. 构建目标文件夹路径
|
||||
# 2. 构建目标文件夹路径 ../data/after_screening/Group/Anion
|
||||
if group_name == anion_name:
|
||||
target_subdir = os.path.join(TARGET_DIR, group_name)
|
||||
else:
|
||||
target_subdir = os.path.join(TARGET_DIR, group_name, anion_name)
|
||||
|
||||
if not os.path.exists(target_subdir):
|
||||
os.makedirs(target_subdir)
|
||||
|
||||
# 3. 构建目标链接路径
|
||||
target_link_path = os.path.join(target_subdir, f"{material_id}.cif")
|
||||
# 3. 构建目标文件路径
|
||||
target_file_path = os.path.join(target_subdir, f"{material_id}.cif")
|
||||
|
||||
# 4. 创建链接
|
||||
# 4. 执行复制 (改为复制以确保结果独立)
|
||||
try:
|
||||
# 如果目标已经存在（可能是旧的链接），先删除
|
||||
if os.path.exists(target_link_path) or os.path.islink(target_link_path):
|
||||
os.remove(target_link_path)
|
||||
if os.path.exists(target_file_path):
|
||||
os.remove(target_file_path)
|
||||
|
||||
shutil.copy(abs_source_path, target_file_path)
|
||||
# 如果你非常确定要软链接，请注释上一行，解开下一行：
|
||||
# os.symlink(abs_source_path, target_file_path)
|
||||
|
||||
os.symlink(abs_source_path, target_link_path)
|
||||
# print(f"Link: {material_id} -> Passed")
|
||||
except OSError as e:
|
||||
print(f"创建软链接失败 {material_id}: {e}")
|
||||
print(f"创建文件失败 {material_id}: {e}")
|
||||
|
||||
|
||||
def process_all_csvs():
|
||||
@@ -101,46 +109,53 @@ def process_all_csvs():
|
||||
|
||||
print("开始执行 Step 2-4 联合筛选...")
|
||||
|
||||
# 遍历 output 目录
|
||||
# 结构预期: ../output/Group/Anion/Anion.csv (例如 ../output/O+S/O/O.csv 或 ../output/O/O.csv)
|
||||
for root, dirs, files in os.walk(CSV_ROOT_DIR):
|
||||
for file in files:
|
||||
if file.endswith(".csv"):
|
||||
csv_path = os.path.join(root, file)
|
||||
|
||||
# 推断 Group 和 Anion
|
||||
# root 的末尾应该是 .../Group/Anion
|
||||
# 例如 root = ../output/O+S/O
|
||||
# --- 核心修正 1: 路径解析逻辑 ---
|
||||
# 获取相对于 output 根目录的路径部分
|
||||
# 例如 root = ../output/S -> rel_root = S
|
||||
# 例如 root = ../output/S+O/S -> rel_root = S+O/S
|
||||
rel_root = os.path.relpath(root, CSV_ROOT_DIR)
|
||||
path_parts = rel_root.split(os.sep)
|
||||
|
||||
path_parts = os.path.normpath(root).split(os.sep)
|
||||
# 倒数第一级是 Anion (O), 倒数第二级是 Group (O+S)
|
||||
if len(path_parts) >= 2:
|
||||
anion_name = path_parts[-1]
|
||||
group_name = path_parts[-2]
|
||||
# 解析 Group 和 Anion
|
||||
if len(path_parts) == 1:
|
||||
# 单层目录: output/S -> Group=S, Anion=S
|
||||
group_name = path_parts[0]
|
||||
anion_name = path_parts[0]
|
||||
elif len(path_parts) >= 2:
|
||||
# 双层目录: output/S+O/S -> Group=S+O, Anion=S
|
||||
group_name = path_parts[0]
|
||||
anion_name = path_parts[1]
|
||||
else:
|
||||
print(f"跳过路径结构异常的 CSV: {csv_path}")
|
||||
# 根目录下直接有csv的情况 (不应该发生)
|
||||
continue
|
||||
|
||||
# 确保这是一个有效的阴离子类型
|
||||
if anion_name not in THRESHOLDS:
|
||||
print(f"跳过不支持的阴离子类型: {anion_name}")
|
||||
continue
|
||||
|
||||
print(f"正在处理: Group={group_name}, Anion={anion_name} ({file})")
|
||||
|
||||
# 读取 CSV
|
||||
df = pd.read_csv(csv_path)
|
||||
# --- 核心修正 2: 防止 Filename 被读取为浮点数 ---
|
||||
# dtypeStr={'Filename': str} 强制将第一列读取为字符串
|
||||
df = pd.read_csv(csv_path, dtype={'Filename': str})
|
||||
|
||||
pass_count = 0
|
||||
total_count = len(df)
|
||||
|
||||
for index, row in df.iterrows():
|
||||
material_id = str(row['Filename'])
|
||||
# 去除可能存在的 .0 后缀 (以防万一 CSV 里已经写成了浮点格式)
|
||||
material_id = str(row['Filename']).replace('.0', '')
|
||||
|
||||
if check_requirements(row, anion_name):
|
||||
create_symlink(group_name, anion_name, material_id)
|
||||
create_result_file(group_name, anion_name, material_id)
|
||||
pass_count += 1
|
||||
|
||||
print(f" - 完成: {pass_count}/{total_count} 个材料通过筛选并建立链接。")
|
||||
print(f" - 完成: {pass_count}/{total_count} 个材料通过筛选并保存至 {TARGET_DIR}。")
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
|
||||
Reference in New Issue
Block a user