修上传BUG
This commit is contained in:
@@ -42,19 +42,30 @@
|
||||
.preview-box {flex: 1; text-align: center; }
|
||||
.preview-box h3 {margin-top: 0;color: #334155; }
|
||||
.preview-box img { max-width: 100%;max-height: 300px;border: 1px solid #e2e8f0;border-radius: 8px;object-fit: contain;}
|
||||
.preview-list {display: grid;grid-template-columns: repeat(auto-fit, minmax(140px, 1fr));gap: 12px;}
|
||||
.preview-list {display: grid;grid-template-columns: repeat(auto-fit, minmax(140px, 1fr));gap: 12px; margin-top: 20px;}
|
||||
.preview-item {position: relative;}
|
||||
.preview-item img {width: 100%;max-height: 220px;border: 1px solid #e2e8f0;border-radius: 8px;object-fit: contain;}
|
||||
.preview-remove {position: absolute;top: 6px;right: 6px;border: none;border-radius: 999px;background: rgba(15,23,42,0.8);color: #fff;width: 24px;height: 24px;cursor: pointer;display: flex;align-items: center;justify-content: center;font-size: 14px;line-height: 1;}
|
||||
.result-box {flex: 1;}
|
||||
.result-box h3 { margin-top: 0; color: #334155;}
|
||||
.form-controls { display: flex;gap: 8px;margin-bottom: 12px;flex-wrap: wrap;}
|
||||
#kvForm {border: 1px solid #e2e8f0; border-radius: 8px; padding: 12px; max-height: 400px; overflow: auto;margin-bottom: 12px;background: #f8fafc;}
|
||||
.form-header { display: grid; grid-template-columns: 1fr 1fr auto; gap: 8px; margin-bottom: 8px; padding: 0 4px; font-weight: 600; color: #475569; font-size: 14px;}
|
||||
.pending-item { background: #fff; border: 1px solid #e2e8f0; border-radius: 12px; padding: 20px; margin-bottom: 24px; box-shadow: 0 4px 6px -1px rgba(0,0,0,0.1); }
|
||||
.pending-item-header { display: flex; justify-content: space-between; align-items: center; margin-bottom: 16px; border-bottom: 1px solid #f1f5f9; padding-bottom: 12px; }
|
||||
.pending-item-title { font-weight: 600; color: #1e293b; font-size: 16px; }
|
||||
.pending-item-body { display: flex; gap: 20px; }
|
||||
.pending-item-preview { flex: 0 0 240px; }
|
||||
.pending-item-preview img { width: 100%; border-radius: 8px; border: 1px solid #f1f5f9; }
|
||||
.pending-item-edit { flex: 1; }
|
||||
.pending-item-footer { margin-top: 16px; text-align: right; }
|
||||
@media (max-width: 992px) {
|
||||
.pending-item-body { flex-direction: column; }
|
||||
.pending-item-preview { flex: 0 0 auto; }
|
||||
}
|
||||
.form-row {display: grid;grid-template-columns: 1fr 1fr auto;gap: 8px; margin-bottom: 6px; align-items: center;}
|
||||
.form-row input {padding: 8px;border: 1px solid #cbd5e1;border-radius: 4px; width: 100%; box-sizing: border-box;}
|
||||
#resultBox { width: 100%;min-height: 200px;font-family: ui-monospace, SFMono-Regular, Menlo, monospace;font-size: 14px; padding: 12px; border: 1px solid #e2e8f0;
|
||||
border-radius: 8px; resize: vertical;box-sizing: border-box; }
|
||||
.kv-form-container {border: 1px solid #e2e8f0; border-radius: 8px; padding: 12px; max-height: 400px; overflow: auto; margin-bottom: 12px; background: #f8fafc;}
|
||||
.form-header { display: grid; grid-template-columns: 1fr 1fr auto; gap: 8px; margin-bottom: 8px; padding: 0 4px; font-weight: 600; color: #475569; font-size: 14px;}
|
||||
.result-textarea { width: 100%; min-height: 120px; font-family: ui-monospace, SFMono-Regular, Menlo, monospace; font-size: 13px; padding: 10px; border: 1px solid #e2e8f0; border-radius: 8px; resize: vertical; box-sizing: border-box; }
|
||||
.status-message { padding: 10px; margin: 10px 0; border-radius: 6px; display: none; }
|
||||
.status-message.success { background-color: #d4edda; color: #155724; border: 1px solid #c3e6cb; }
|
||||
.status-message.error { background-color: #f8d7da; color: #721c24; border: 1px solid #f5c6cb; }
|
||||
@@ -96,6 +107,7 @@
|
||||
{% csrf_token %}
|
||||
<input type="file" id="fileInput" name="file" accept="image/*,.pdf" multiple />
|
||||
<span id="fileHint" class="muted"></span>
|
||||
<div id="previewList" class="preview-list"></div>
|
||||
<br>
|
||||
<button type="submit" class="btn btn-primary">上传并识别</button>
|
||||
</form>
|
||||
@@ -107,19 +119,11 @@
|
||||
</div>
|
||||
|
||||
<div class="preview-container">
|
||||
<div class="preview-box">
|
||||
<h3>图片预览</h3>
|
||||
<div id="previewList" class="preview-list"></div>
|
||||
</div>
|
||||
|
||||
<div class="result-box">
|
||||
<h3>识别结果(可编辑)</h3>
|
||||
<div class="form-controls">
|
||||
<button id="addFieldBtn" class="btn btn-secondary" type="button">添加字段</button>
|
||||
<button id="syncFromTextBtn" class="btn btn-secondary" type="button">从文本区刷新表单</button>
|
||||
<h3>待处理文件列表</h3>
|
||||
<div id="pendingItems" class="pending-list">
|
||||
<!-- 这里将动态生成每个文件的预览和编辑区域 -->
|
||||
</div>
|
||||
<div id="kvForm"></div>
|
||||
<textarea id="resultBox" placeholder="识别结果JSON将显示在这里"></textarea>
|
||||
</div>
|
||||
</div>
|
||||
|
||||
@@ -142,20 +146,17 @@ const uploadForm = document.getElementById('uploadForm');
|
||||
const fileInput = document.getElementById('fileInput');
|
||||
const fileHint = document.getElementById('fileHint');
|
||||
const previewList = document.getElementById('previewList');
|
||||
const resultBox = document.getElementById('resultBox');
|
||||
const pendingItems = document.getElementById('pendingItems');
|
||||
const uploadMsg = document.getElementById('uploadMsg');
|
||||
const confirmBtn = document.getElementById('confirmBtn');
|
||||
const clearBtn = document.getElementById('clearBtn');
|
||||
const confirmMsg = document.getElementById('confirmMsg');
|
||||
const kvForm = document.getElementById('kvForm');
|
||||
const addFieldBtn = document.getElementById('addFieldBtn');
|
||||
const syncFromTextBtn = document.getElementById('syncFromTextBtn');
|
||||
const dropArea = document.getElementById('dropArea');
|
||||
const progressWrap = document.getElementById('progressWrap');
|
||||
const progressBar = document.getElementById('progressBar');
|
||||
const progressText = document.getElementById('progressText');
|
||||
|
||||
let currentImageRel = [];
|
||||
let currentItems = []; // 存储当前待处理的所有文件结果
|
||||
let selectedFiles = [];
|
||||
|
||||
function setProgress(p, text){
|
||||
@@ -251,10 +252,15 @@ function setPreviewList(urls) {
|
||||
const idx = Number(item.dataset.index);
|
||||
if (!Number.isNaN(idx)) {
|
||||
selectedFiles.splice(idx, 1);
|
||||
const urls = selectedFiles.map(f => URL.createObjectURL(f));
|
||||
const urls = selectedFiles.map(f => {
|
||||
if (f.name.toLowerCase().endsWith('.pdf')) {
|
||||
return 'data:image/svg+xml;base64,PHN2ZyB4bWxucz0iaHR0cDovL3d3dy53My5vcmcvMjAwMC9zdmciIHdpZHRoPSI0OCIgaGVpZ2h0PSI0OCIgdmlld0JveD0iMCAwIDI0IDI0IiBmaWxsPSJub25lIiBzdHJva2U9IiNlZjQ0NDQiIHN0cm9rZS13aWR0aD0iMiIgc3Ryb2tlLWxpbmVjYXA9InJvdW5kIiBzdHJva2UtbGluZWpvaW49InJvdW5kIj48cGF0aCBkPSJNMTQgMmgyYTIgMiAwIDAgMSAyIDJ2MTZhMiAyIDAgMCAxLTIgMmgtMTJhMiAyIDAgMCAxLTItMlY0YTIgMiAwIDAgMSAyLTJoMiIvPjxwYXRoIGQ9Ik0xNCAydjRjMCAxLjEgLjkgMiAyIDJoNCIvPjxwYXRoIGQ9Ik03IDloNSIvPjxwYXRoIGQ9Ik03IDEzaDUiLz48cGF0aCBkPSJNNyAxN2g4Ii8+PC9zdmc+';
|
||||
}
|
||||
return URL.createObjectURL(f);
|
||||
});
|
||||
setPreviewList(urls);
|
||||
updateFileHint();
|
||||
setTimeout(() => urls.forEach(u => URL.revokeObjectURL(u)), 0);
|
||||
setTimeout(() => urls.forEach(u => { if (u.startsWith('blob:')) URL.revokeObjectURL(u); }), 0);
|
||||
}
|
||||
};
|
||||
item.appendChild(img);
|
||||
@@ -265,7 +271,7 @@ function setPreviewList(urls) {
|
||||
|
||||
function updateFileHint() {
|
||||
const count = selectedFiles.length;
|
||||
fileHint.textContent = count ? `已选择 ${count} 张` : '未选择文件';
|
||||
fileHint.textContent = count ? `已选择 ${count} 个文件` : '未选择文件';
|
||||
}
|
||||
|
||||
function addFiles(files) {
|
||||
@@ -280,18 +286,13 @@ function addFiles(files) {
|
||||
});
|
||||
const urls = selectedFiles.map(f => {
|
||||
if (f.name.toLowerCase().endsWith('.pdf')) {
|
||||
// 使用一个简单的 SVG PDF 图标 Data URI
|
||||
return 'data:image/svg+xml;base64,PHN2ZyB4bWxucz0iaHR0cDovL3d3dy53My5vcmcvMjAwMC9zdmciIHdpZHRoPSI0OCIgaGVpZ2h0PSI0OCIgdmlld0JveD0iMCAwIDI0IDI0IiBmaWxsPSJub25lIiBzdHJva2U9IiNlZjQ0NDQiIHN0cm9rZS13aWR0aD0iMiIgc3Ryb2tlLWxpbmVjYXA9InJvdW5kIiBzdHJva2UtbGluZWpvaW49InJvdW5kIj48cGF0aCBkPSJNMTQgMmgyYTIgMiAwIDAgMSAyIDJ2MTZhMiAyIDAgMCAxLTIgMmgtMTJhMiAyIDAgMCAxLTItMlY0YTIgMiAwIDAgMSAyLTJoMiIvPjxwYXRoIGQ9Ik0xNCAydjRjMCAxLjEgLjkgMiAyIDJoNCIvPjxwYXRoIGQ9Ik03IDloNSIvPjxwYXRoIGQ9Ik03IDEzaDUiLz48cGF0aCBkPSJNNyAxN2g4Ii8+PC9zdmc+';
|
||||
}
|
||||
return URL.createObjectURL(f);
|
||||
});
|
||||
setPreviewList(urls);
|
||||
updateFileHint();
|
||||
setTimeout(() => {
|
||||
urls.forEach(u => {
|
||||
if (u.startsWith('blob:')) URL.revokeObjectURL(u);
|
||||
});
|
||||
}, 0);
|
||||
setTimeout(() => urls.forEach(u => { if (u.startsWith('blob:')) URL.revokeObjectURL(u); }), 0);
|
||||
}
|
||||
|
||||
fileInput.addEventListener('change', function(e) {
|
||||
@@ -299,7 +300,7 @@ fileInput.addEventListener('change', function(e) {
|
||||
fileInput.value = '';
|
||||
});
|
||||
|
||||
function createRow(k = '', v = '') {
|
||||
function createKvRow(k = '', v = '', onInput) {
|
||||
const row = document.createElement('div');
|
||||
row.className = 'form-row';
|
||||
const keyInput = document.createElement('input');
|
||||
@@ -314,93 +315,147 @@ function createRow(k = '', v = '') {
|
||||
delBtn.type = 'button';
|
||||
delBtn.className = 'btn btn-danger';
|
||||
delBtn.textContent = '删除';
|
||||
|
||||
delBtn.onclick = () => {
|
||||
if (kvForm.querySelectorAll('.form-row').length > 1) {
|
||||
kvForm.removeChild(row);
|
||||
const container = row.parentElement;
|
||||
if (container.querySelectorAll('.form-row').length > 1) {
|
||||
container.removeChild(row);
|
||||
} else {
|
||||
keyInput.value = '';
|
||||
valInput.value = '';
|
||||
}
|
||||
syncTextarea();
|
||||
if (onInput) onInput();
|
||||
};
|
||||
keyInput.oninput = syncTextarea;
|
||||
valInput.oninput = syncTextarea;
|
||||
|
||||
keyInput.oninput = onInput;
|
||||
valInput.oninput = onInput;
|
||||
|
||||
row.appendChild(keyInput);
|
||||
row.appendChild(valInput);
|
||||
row.appendChild(delBtn);
|
||||
return row;
|
||||
}
|
||||
|
||||
function renderFormFromObject(obj) {
|
||||
kvForm.innerHTML = `
|
||||
<div class="form-header">
|
||||
<div>字段名</div>
|
||||
<div>字段值</div>
|
||||
<div>操作</div>
|
||||
</div>
|
||||
`;
|
||||
Object.keys(obj || {}).forEach(k => {
|
||||
kvForm.appendChild(createRow(k, obj[k]));
|
||||
});
|
||||
if (kvForm.children.length <= 1) kvForm.appendChild(createRow());
|
||||
syncTextarea();
|
||||
function renderPendingItems(items) {
|
||||
pendingItems.innerHTML = '';
|
||||
currentItems = items;
|
||||
|
||||
items.forEach((item, index) => {
|
||||
const itemEl = document.createElement('div');
|
||||
itemEl.className = 'pending-item';
|
||||
|
||||
const header = document.createElement('div');
|
||||
header.className = 'pending-item-header';
|
||||
header.innerHTML = `<span class="pending-item-title">${index + 1}. ${item.name}</span>`;
|
||||
|
||||
const removeBtn = document.createElement('button');
|
||||
removeBtn.className = 'btn btn-danger';
|
||||
removeBtn.textContent = '忽略此项';
|
||||
removeBtn.onclick = () => {
|
||||
currentItems.splice(index, 1);
|
||||
renderPendingItems(currentItems);
|
||||
};
|
||||
header.appendChild(removeBtn);
|
||||
|
||||
const body = document.createElement('div');
|
||||
body.className = 'pending-item-body';
|
||||
|
||||
const preview = document.createElement('div');
|
||||
preview.className = 'pending-item-preview';
|
||||
const mainImg = document.createElement('img');
|
||||
mainImg.src = item.image_urls[0];
|
||||
preview.appendChild(mainImg);
|
||||
if (item.image_urls.length > 1) {
|
||||
const hint = document.createElement('p');
|
||||
hint.className = 'muted';
|
||||
hint.style.textAlign = 'center';
|
||||
hint.textContent = `共 ${item.image_urls.length} 页`;
|
||||
preview.appendChild(hint);
|
||||
}
|
||||
|
||||
function objectFromForm() {
|
||||
const edit = document.createElement('div');
|
||||
edit.className = 'pending-item-edit';
|
||||
|
||||
const controls = document.createElement('div');
|
||||
controls.className = 'form-controls';
|
||||
const addBtn = document.createElement('button');
|
||||
addBtn.className = 'btn btn-secondary';
|
||||
addBtn.textContent = '添加字段';
|
||||
const syncBtn = document.createElement('button');
|
||||
syncBtn.className = 'btn btn-secondary';
|
||||
syncBtn.textContent = '刷新表单';
|
||||
controls.appendChild(addBtn);
|
||||
controls.appendChild(syncBtn);
|
||||
|
||||
const kvForm = document.createElement('div');
|
||||
kvForm.className = 'kv-form-container';
|
||||
kvForm.innerHTML = '<div class="form-header"><div>字段名</div><div>字段值</div><div>操作</div></div>';
|
||||
|
||||
const textarea = document.createElement('textarea');
|
||||
textarea.className = 'result-textarea';
|
||||
|
||||
const syncData = () => {
|
||||
const obj = {};
|
||||
Array.from(kvForm.children).forEach(row => {
|
||||
if (row.classList.contains('form-header')) return;
|
||||
const [kInput, vInput] = row.querySelectorAll('input');
|
||||
const k = (kInput.value || '').trim();
|
||||
kvForm.querySelectorAll('.form-row').forEach(row => {
|
||||
const inputs = row.querySelectorAll('input');
|
||||
const k = inputs[0].value.trim();
|
||||
if (!k) return;
|
||||
const raw = vInput.value;
|
||||
try { obj[k] = JSON.parse(inputs[1].value); } catch(e) { obj[k] = inputs[1].value; }
|
||||
});
|
||||
item.data = obj;
|
||||
textarea.value = JSON.stringify(obj, null, 2);
|
||||
};
|
||||
|
||||
Object.entries(item.data).forEach(([k, v]) => {
|
||||
kvForm.appendChild(createKvRow(k, v, syncData));
|
||||
});
|
||||
if (kvForm.querySelectorAll('.form-row').length === 0) {
|
||||
kvForm.appendChild(createKvRow('', '', syncData));
|
||||
}
|
||||
|
||||
addBtn.onclick = () => {
|
||||
kvForm.appendChild(createKvRow('', '', syncData));
|
||||
syncData();
|
||||
};
|
||||
|
||||
syncBtn.onclick = () => {
|
||||
try {
|
||||
obj[k] = JSON.parse(raw);
|
||||
} catch (_) {
|
||||
obj[k] = raw;
|
||||
}
|
||||
});
|
||||
return obj;
|
||||
}
|
||||
const obj = JSON.parse(textarea.value);
|
||||
kvForm.innerHTML = '<div class="form-header"><div>字段名</div><div>字段值</div><div>操作</div></div>';
|
||||
Object.entries(obj).forEach(([k, v]) => kvForm.appendChild(createKvRow(k, v, syncData)));
|
||||
item.data = obj;
|
||||
} catch(e) { alert('JSON格式错误'); }
|
||||
};
|
||||
|
||||
function syncTextarea() {
|
||||
const obj = objectFromForm();
|
||||
resultBox.value = JSON.stringify(obj, null, 2);
|
||||
}
|
||||
textarea.value = JSON.stringify(item.data, null, 2);
|
||||
textarea.oninput = () => { item.data = JSON.parse(textarea.value); };
|
||||
|
||||
addFieldBtn.addEventListener('click', () => {
|
||||
kvForm.appendChild(createRow());
|
||||
syncTextarea();
|
||||
edit.appendChild(controls);
|
||||
edit.appendChild(kvForm);
|
||||
edit.appendChild(textarea);
|
||||
|
||||
body.appendChild(preview);
|
||||
body.appendChild(edit);
|
||||
|
||||
itemEl.appendChild(header);
|
||||
itemEl.appendChild(body);
|
||||
pendingItems.appendChild(itemEl);
|
||||
});
|
||||
|
||||
syncFromTextBtn.addEventListener('click', () => {
|
||||
try {
|
||||
const obj = JSON.parse(resultBox.value || '{}');
|
||||
renderFormFromObject(obj);
|
||||
uploadMsg.textContent = '已从文本区刷新表单';
|
||||
uploadMsg.className = 'status-message success';
|
||||
uploadMsg.style.display = 'block';
|
||||
setTimeout(() => {
|
||||
uploadMsg.style.display = 'none';
|
||||
}, 2000);
|
||||
} catch (e) {
|
||||
uploadMsg.textContent = '文本区不是有效JSON';
|
||||
uploadMsg.className = 'status-message error';
|
||||
uploadMsg.style.display = 'block';
|
||||
confirmBtn.disabled = items.length === 0;
|
||||
}
|
||||
});
|
||||
|
||||
uploadForm.addEventListener('submit', async (e) => {
|
||||
e.preventDefault();
|
||||
uploadMsg.textContent = '';
|
||||
confirmMsg.textContent = '';
|
||||
confirmBtn.disabled = true;
|
||||
resultBox.value = '';
|
||||
currentImageRel = [];
|
||||
previewList.innerHTML = '';
|
||||
pendingItems.innerHTML = '';
|
||||
currentItems = [];
|
||||
|
||||
const files = Array.from(selectedFiles || []).filter(f => f && (f.type.startsWith('image/') || f.name.toLowerCase().endsWith('.pdf')));
|
||||
if (!files.length) {
|
||||
uploadMsg.textContent = '请选择图片或PDF文件';
|
||||
if (!selectedFiles.length) {
|
||||
uploadMsg.textContent = '请选择文件';
|
||||
uploadMsg.className = 'status-message error';
|
||||
uploadMsg.style.display = 'block';
|
||||
return;
|
||||
@@ -409,10 +464,10 @@ uploadForm.addEventListener('submit', async (e) => {
|
||||
showProgress();
|
||||
setProgress(5, '预处理中');
|
||||
const formData = new FormData();
|
||||
for (let i = 0; i < files.length; i++) {
|
||||
const file = files[i];
|
||||
for (let i = 0; i < selectedFiles.length; i++) {
|
||||
const file = selectedFiles[i];
|
||||
if (file.type.startsWith('image/')) {
|
||||
setProgress(5 + Math.round((i/files.length)*45), '转换图片');
|
||||
setProgress(5 + Math.round((i/selectedFiles.length)*45), '转换图片');
|
||||
try {
|
||||
const jpegFile = await convertToJpeg(file);
|
||||
formData.append('file', jpegFile);
|
||||
@@ -420,7 +475,6 @@ uploadForm.addEventListener('submit', async (e) => {
|
||||
formData.append('file', file);
|
||||
}
|
||||
} else {
|
||||
// PDF 直接添加
|
||||
formData.append('file', file);
|
||||
}
|
||||
}
|
||||
@@ -431,7 +485,8 @@ uploadForm.addEventListener('submit', async (e) => {
|
||||
const timer = setInterval(() => {
|
||||
prog = Math.min(95, prog + 1);
|
||||
setProgress(prog, '识别中');
|
||||
}, 120);
|
||||
}, 200);
|
||||
|
||||
const resp = await fetch('/elastic/upload/', {
|
||||
method: 'POST',
|
||||
credentials: 'same-origin',
|
||||
@@ -447,11 +502,8 @@ uploadForm.addEventListener('submit', async (e) => {
|
||||
uploadMsg.textContent = data.message || '识别成功';
|
||||
uploadMsg.className = 'status-message success';
|
||||
uploadMsg.style.display = 'block';
|
||||
const urls = data.image_urls || (data.image_url ? [data.image_url] : []);
|
||||
setPreviewList(urls);
|
||||
renderFormFromObject(data.data || {});
|
||||
currentImageRel = data.images || (data.image ? [data.image] : []);
|
||||
confirmBtn.disabled = false;
|
||||
|
||||
renderPendingItems(data.items || []);
|
||||
setTimeout(hideProgress, 800);
|
||||
} catch (e) {
|
||||
uploadMsg.textContent = e.message || '发生错误';
|
||||
@@ -462,9 +514,14 @@ uploadForm.addEventListener('submit', async (e) => {
|
||||
});
|
||||
|
||||
confirmBtn.addEventListener('click', async () => {
|
||||
confirmMsg.textContent = '';
|
||||
confirmMsg.textContent = '正在录入...';
|
||||
try {
|
||||
const edited = objectFromForm();
|
||||
const payload = {
|
||||
items: currentItems.map(it => ({
|
||||
data: it.data,
|
||||
image: it.images
|
||||
}))
|
||||
};
|
||||
const resp = await fetch('/elastic/confirm/', {
|
||||
method: 'POST',
|
||||
credentials: 'same-origin',
|
||||
@@ -472,7 +529,7 @@ confirmBtn.addEventListener('click', async () => {
|
||||
'Content-Type': 'application/json',
|
||||
'X-CSRFToken': getCookie('csrftoken') || ''
|
||||
},
|
||||
body: JSON.stringify({ data: edited, image: currentImageRel })
|
||||
body: JSON.stringify(payload)
|
||||
});
|
||||
const data = await resp.json();
|
||||
if (!resp.ok || data.status !== 'success') {
|
||||
@@ -480,6 +537,12 @@ confirmBtn.addEventListener('click', async () => {
|
||||
}
|
||||
confirmMsg.textContent = data.message || '录入成功';
|
||||
confirmMsg.style.color = '#179957';
|
||||
// 录入成功后清空待处理列表
|
||||
pendingItems.innerHTML = '';
|
||||
currentItems = [];
|
||||
selectedFiles = [];
|
||||
updateFileHint();
|
||||
confirmBtn.disabled = true;
|
||||
} catch (e) {
|
||||
confirmMsg.textContent = e.message || '发生错误';
|
||||
confirmMsg.style.color = '#d14343';
|
||||
@@ -489,13 +552,11 @@ confirmBtn.addEventListener('click', async () => {
|
||||
clearBtn.addEventListener('click', () => {
|
||||
fileInput.value = '';
|
||||
previewList.innerHTML = '';
|
||||
resultBox.value = '';
|
||||
kvForm.innerHTML = '';
|
||||
kvForm.appendChild(createRow()); // 保留一个空行
|
||||
pendingItems.innerHTML = '';
|
||||
uploadMsg.textContent = '';
|
||||
confirmMsg.textContent = '';
|
||||
confirmBtn.disabled = true;
|
||||
currentImageRel = [];
|
||||
currentItems = [];
|
||||
selectedFiles = [];
|
||||
updateFileHint();
|
||||
});
|
||||
|
||||
174
elastic/views.py
174
elastic/views.py
@@ -607,107 +607,96 @@ def upload(request):
|
||||
|
||||
images_dir = os.path.join(settings.MEDIA_ROOT, "images")
|
||||
os.makedirs(images_dir, exist_ok=True)
|
||||
rel_paths = []
|
||||
image_urls = []
|
||||
data_list = []
|
||||
|
||||
# 预处理文件列表,处理PDF转换
|
||||
processed_files = []
|
||||
# 按照原始文件进行分组处理
|
||||
file_results = []
|
||||
|
||||
for f in files:
|
||||
if f.name.lower().endswith('.pdf'):
|
||||
group_images = [] # 存储该文件生成的所有图片路径信息 (abs_path, filename)
|
||||
is_pdf = f.name.lower().endswith('.pdf')
|
||||
|
||||
if is_pdf:
|
||||
if not HAS_PDF_SUPPORT:
|
||||
return JsonResponse({"status": "error", "message": f"服务器未安装PDF处理组件(PyMuPDF): {PDF_ERROR}"}, status=500)
|
||||
|
||||
# 将PDF保存到临时文件
|
||||
with tempfile.NamedTemporaryFile(delete=False, suffix='.pdf') as tmp:
|
||||
for chunk in f.chunks():
|
||||
tmp.write(chunk)
|
||||
tmp_path = tmp.name
|
||||
|
||||
try:
|
||||
# 转换PDF为图片列表
|
||||
doc = fitz.open(tmp_path)
|
||||
for i in range(len(doc)):
|
||||
page = doc.load_page(i)
|
||||
# 降低 DPI 从 200 到 150,在保证准确率的同时显著减小图片体积,加快上传速度
|
||||
pix = page.get_pixmap(dpi=150)
|
||||
img_filename = f"{uuid.uuid4()}_page_{i+1}.jpg"
|
||||
img_abs_path = os.path.join(images_dir, img_filename)
|
||||
pix.save(img_abs_path)
|
||||
processed_files.append((img_abs_path, img_filename, f"{f.name}_p{i+1}"))
|
||||
group_images.append((img_abs_path, img_filename))
|
||||
doc.close()
|
||||
except Exception as e:
|
||||
return JsonResponse({"status": "error", "message": f"PDF转换失败: {str(e)}"}, status=500)
|
||||
return JsonResponse({"status": "error", "message": f"PDF {f.name} 转换失败: {str(e)}"}, status=500)
|
||||
finally:
|
||||
if os.path.exists(tmp_path):
|
||||
os.remove(tmp_path)
|
||||
else:
|
||||
# 普通图片处理
|
||||
filename = f"{uuid.uuid4()}_{f.name}"
|
||||
abs_path = os.path.join(images_dir, filename)
|
||||
with open(abs_path, "wb") as dst:
|
||||
for chunk in f.chunks():
|
||||
dst.write(chunk)
|
||||
processed_files.append((abs_path, filename, f.name))
|
||||
group_images.append((abs_path, filename))
|
||||
|
||||
# 使用线程池并行调用 OCR 接口以提升速度
|
||||
def run_ocr(file_info):
|
||||
abs_path, filename, original_name = file_info
|
||||
# 对该组图片并行进行 OCR 识别
|
||||
def run_ocr(img_info):
|
||||
abs_p, fname = img_info
|
||||
try:
|
||||
data = ocr_and_extract_info(abs_path)
|
||||
return (data, filename)
|
||||
except Exception as e:
|
||||
return (e, filename)
|
||||
data = ocr_and_extract_info(abs_p)
|
||||
return data
|
||||
except Exception:
|
||||
return None
|
||||
|
||||
rel_paths = []
|
||||
image_urls = []
|
||||
data_list = []
|
||||
group_data_list = []
|
||||
with concurrent.futures.ThreadPoolExecutor(max_workers=min(len(group_images), 8)) as executor:
|
||||
futures = [executor.submit(run_ocr, img_info) for img_info in group_images]
|
||||
for future in concurrent.futures.as_completed(futures):
|
||||
res = future.result()
|
||||
if res:
|
||||
group_data_list.append(res)
|
||||
|
||||
# 限制最大线程数为 8,避免过多并发导致 API 限制或资源耗尽
|
||||
with concurrent.futures.ThreadPoolExecutor(max_workers=8) as executor:
|
||||
future_to_file = {executor.submit(run_ocr, f_info): f_info for f_info in processed_files}
|
||||
for future in concurrent.futures.as_completed(future_to_file):
|
||||
result, filename = future.result()
|
||||
if isinstance(result, Exception):
|
||||
return JsonResponse({"status": "error", "message": f"文件 {filename} 识别失败: {str(result)}"}, status=500)
|
||||
|
||||
if result:
|
||||
data_list.append(result)
|
||||
|
||||
rel_path = f"images/{filename}"
|
||||
rel_paths.append(rel_path)
|
||||
image_urls.append(request.build_absolute_uri(settings.MEDIA_URL + rel_path))
|
||||
|
||||
if not data_list:
|
||||
return JsonResponse({"status": "error", "message": "无法识别图片内容"}, status=400)
|
||||
|
||||
merged = {}
|
||||
for item in data_list:
|
||||
if not isinstance(item, dict):
|
||||
continue
|
||||
# 合并该文件的多页识别结果
|
||||
merged_group_data = {}
|
||||
for item in group_data_list:
|
||||
if not isinstance(item, dict): continue
|
||||
for k, v in item.items():
|
||||
key = str(k).strip()
|
||||
if not key:
|
||||
continue
|
||||
if key not in merged or merged.get(key) in (None, ''):
|
||||
merged[key] = v
|
||||
continue
|
||||
if merged.get(key) == v:
|
||||
continue
|
||||
if not key: continue
|
||||
if key not in merged_group_data or merged_group_data.get(key) in (None, ''):
|
||||
merged_group_data[key] = v
|
||||
elif merged_group_data.get(key) != v:
|
||||
base = key
|
||||
idx = 2
|
||||
while f"{base}_{idx}" in merged:
|
||||
idx += 1
|
||||
merged[f"{base}_{idx}"] = v
|
||||
while f"{base}_{idx}" in merged_group_data: idx += 1
|
||||
merged_group_data[f"{base}_{idx}"] = v
|
||||
|
||||
if not merged_group_data:
|
||||
# 如果没识别到,至少保留一个空结构或者包含文件名的提示
|
||||
merged_group_data = {"文件名": f.name, "提示": "未识别到具体内容"}
|
||||
|
||||
rel_paths = [f"images/{img[1]}" for img in group_images]
|
||||
image_urls = [request.build_absolute_uri(settings.MEDIA_URL + rp) for rp in rel_paths]
|
||||
|
||||
file_results.append({
|
||||
"name": f.name,
|
||||
"data": merged_group_data,
|
||||
"images": rel_paths,
|
||||
"image_urls": image_urls,
|
||||
})
|
||||
|
||||
return JsonResponse({
|
||||
"status": "success",
|
||||
"message": "识别成功,请确认数据后点击录入",
|
||||
"data": merged,
|
||||
"images": rel_paths,
|
||||
"image_urls": image_urls,
|
||||
"image": rel_paths[0] if rel_paths else "",
|
||||
"image_url": image_urls[0] if image_urls else "",
|
||||
"message": f"成功处理 {len(file_results)} 个文件,请确认数据后点击录入",
|
||||
"items": file_results,
|
||||
})
|
||||
|
||||
|
||||
@@ -715,7 +704,6 @@ def upload(request):
|
||||
@require_http_methods(["POST"])
|
||||
def confirm(request):
|
||||
if request.session.get("user_id") is None:
|
||||
# 允许从payload中带入user_id作为后备(便于前端已知用户时继续操作)
|
||||
try:
|
||||
payload_for_uid = json.loads(request.body.decode("utf-8"))
|
||||
except Exception:
|
||||
@@ -732,15 +720,31 @@ def confirm(request):
|
||||
except json.JSONDecodeError:
|
||||
return JsonResponse({"status": "error", "message": "JSON无效"}, status=400)
|
||||
|
||||
edited = payload.get("data") or {}
|
||||
image_rel = payload.get("image") or ""
|
||||
# 支持单项或批量入库
|
||||
items = payload.get("items")
|
||||
if not items:
|
||||
# 兼容旧版本单项入库
|
||||
items = [{
|
||||
"data": payload.get("data"),
|
||||
"image": payload.get("image")
|
||||
}]
|
||||
|
||||
success_count = 0
|
||||
errors = []
|
||||
|
||||
for item in items:
|
||||
edited = item.get("data") or {}
|
||||
image_rel = item.get("image") or ""
|
||||
|
||||
if not isinstance(edited, dict) or not edited:
|
||||
return JsonResponse({"status": "error", "message": "数据不能为空"}, status=400)
|
||||
errors.append("数据项不能为空")
|
||||
continue
|
||||
|
||||
ensure_type_in_list(edited.get("数据类型"))
|
||||
image_ref_to_store = ""
|
||||
temp_files_to_delete = []
|
||||
image_rels = _parse_image_refs(image_rel)
|
||||
|
||||
if image_rels:
|
||||
images_dir = os.path.join(settings.MEDIA_ROOT, "images")
|
||||
os.makedirs(images_dir, exist_ok=True)
|
||||
@@ -748,7 +752,9 @@ def confirm(request):
|
||||
for rel in image_rels:
|
||||
src_abs = os.path.join(settings.MEDIA_ROOT, rel)
|
||||
if not os.path.isfile(src_abs):
|
||||
return JsonResponse({"status": "error", "message": "图片文件不存在"}, status=400)
|
||||
errors.append(f"图片文件 {rel} 不存在")
|
||||
continue
|
||||
|
||||
webp_name = f"{uuid.uuid4().hex}.webp"
|
||||
webp_abs = os.path.join(images_dir, webp_name)
|
||||
try:
|
||||
@@ -759,12 +765,8 @@ def confirm(request):
|
||||
im = im.convert("RGB")
|
||||
im.save(webp_abs, format="WEBP", quality=80)
|
||||
except Exception:
|
||||
try:
|
||||
if os.path.isfile(webp_abs):
|
||||
os.remove(webp_abs)
|
||||
except Exception:
|
||||
pass
|
||||
return JsonResponse({"status": "error", "message": "图片转换WEBP失败"}, status=500)
|
||||
errors.append(f"图片 {rel} 转换WEBP失败")
|
||||
continue
|
||||
|
||||
try:
|
||||
object_name = f"images/{webp_name}"
|
||||
@@ -774,12 +776,12 @@ def confirm(request):
|
||||
image_refs.append(f"minio:{object_name}")
|
||||
temp_files_to_delete.extend([src_abs, webp_abs])
|
||||
else:
|
||||
# Fallback to local storage
|
||||
image_refs.append(f"local:{object_name}")
|
||||
# In local case, we keep the webp file and only delete the original temporary file
|
||||
temp_files_to_delete.append(src_abs)
|
||||
except Exception as e:
|
||||
return JsonResponse({"status": "error", "message": f"存储图片失败: {e}"}, status=500)
|
||||
errors.append(f"存储图片 {rel} 失败: {e}")
|
||||
continue
|
||||
|
||||
if len(image_refs) == 1:
|
||||
image_ref_to_store = image_refs[0]
|
||||
elif len(image_refs) > 1:
|
||||
@@ -792,17 +794,23 @@ def confirm(request):
|
||||
}
|
||||
|
||||
ok = insert_data(to_store)
|
||||
if not ok:
|
||||
return JsonResponse({"status": "error", "message": "写入ES失败"}, status=500)
|
||||
|
||||
try:
|
||||
if ok:
|
||||
success_count += 1
|
||||
# 清理临时文件
|
||||
for p in temp_files_to_delete:
|
||||
if p and os.path.isfile(p):
|
||||
os.remove(p)
|
||||
except Exception:
|
||||
pass
|
||||
try: os.remove(p)
|
||||
except: pass
|
||||
else:
|
||||
errors.append("写入ES失败")
|
||||
|
||||
return JsonResponse({"status": "success", "message": "数据录入成功", "data": edited})
|
||||
if success_count > 0:
|
||||
msg = f"成功录入 {success_count} 条数据"
|
||||
if errors:
|
||||
msg += f" (遇到 {len(errors)} 个错误)"
|
||||
return JsonResponse({"status": "success", "message": msg})
|
||||
else:
|
||||
return JsonResponse({"status": "error", "message": "录入失败: " + "; ".join(errors[:3])}, status=500)
|
||||
|
||||
|
||||
@require_http_methods(["GET"])
|
||||
|
||||
Reference in New Issue
Block a user